Copy disabled (too large)
Download .txt
Showing preview only (11,118K chars total). Download the full file to get everything.
Repository: aws/aws-neuron-sdk
Branch: master
Commit: 371eabc8a739
Files: 1636
Total size: 10.3 MB
Directory structure:
gitextract_u554eb9v/
├── .github/
│ ├── ISSUE_TEMPLATE/
│ │ ├── bug-report.yml
│ │ ├── config.yml
│ │ ├── documentation.yml
│ │ └── feature-request.yml
│ ├── pull_request_template.md
│ ├── stale_issue_mark_close_workflow.yml
│ └── workflows/
│ ├── acknowledge-new-issue.yml
│ └── auto-label-issues.yml
├── .gitignore
├── .readthedocs.yml
├── CODEOWNERS
├── CONTRIBUTING.md
├── Dockerfile
├── LICENSE-DOCUMENTATION
├── LICENSE-SAMPLECODE
├── LICENSE-SUMMARY-DOCS-SAMPLES
├── Makefile
├── README.md
├── _backup-setup/
│ └── neuron-setup/
│ ├── multiframework/
│ │ ├── multi-framework-ubuntu22-neuron-dlami.rst
│ │ └── multi-framework-ubuntu24-neuron-dlami.rst
│ └── pytorch/
│ ├── neuron/
│ │ ├── amazon-linux/
│ │ │ ├── torch-neuron-al2-base-dlami.rst
│ │ │ ├── torch-neuron-al2-pytorch-dlami.rst
│ │ │ ├── torch-neuron-al2.rst
│ │ │ └── torch-neuron-al2023.rst
│ │ └── ubuntu/
│ │ ├── torch-neuron-ubuntu20-base-dlami.rst
│ │ ├── torch-neuron-ubuntu20-pytorch-dlami.rst
│ │ ├── torch-neuron-ubuntu20.rst
│ │ └── torch-neuron-ubuntu22.rst
│ └── neuronx/
│ ├── amazon-linux/
│ │ ├── torch-neuronx-al2-base-dlami.rst
│ │ ├── torch-neuronx-al2-pytorch-dlami.rst
│ │ ├── torch-neuronx-al2.rst
│ │ └── torch-neuronx-al2023.rst
│ └── ubuntu/
│ ├── torch-neuronx-ubuntu20-base-dlami.rst
│ ├── torch-neuronx-ubuntu20-pytorch-dlami.rst
│ ├── torch-neuronx-ubuntu20.rst
│ ├── torch-neuronx-ubuntu22.rst
│ └── torch-neuronx-ubuntu24.rst
├── _content-types/
│ ├── conceptual-deep-dive.rst
│ ├── model-card.rst
│ ├── procedural-how-to.rst
│ ├── procedural-tutorial.ipynb
│ ├── reference-kernel-api.rst
│ └── release-notes-templates/
│ ├── compiler.rst
│ ├── containers.rst
│ ├── dlami.rst
│ ├── index.rst
│ ├── nki.rst
│ ├── nx-jax.rst
│ ├── nx-pytorch.rst
│ ├── nxd-core.rst
│ ├── nxd-inference.rst
│ ├── nxd-training.rst
│ ├── runtime.rst
│ └── tools.rst
├── _ext/
│ ├── archive.py
│ ├── df_tables.py
│ ├── local_documenter.py
│ ├── neuron_tag.py
│ ├── release-notes-automation-spec.md
│ ├── release-notes-context.md
│ ├── sphinx_plotly_directive.py
│ └── symlink.py
├── _static/
│ └── css/
│ ├── custom.css
│ └── custom.css.new
├── _templates/
│ ├── recentposts.html
│ ├── search-field.html
│ ├── search-google.html
│ └── search.html
├── _utilities/
│ ├── JIRA_SETUP_QUICKSTART.md
│ ├── add_meta.py
│ ├── audit_frameworks.py
│ ├── check_urls.sh
│ ├── create_sitemap.py
│ ├── format_build_logs.py
│ ├── inject_archive_meta.py
│ ├── metadata_schema.yaml
│ ├── migrate_setup_content.py
│ ├── old-nki-apis.txt
│ └── setup_jira_token.sh
├── about-neuron/
│ ├── amazonq-getstarted.rst
│ ├── announcements/
│ │ ├── index.rst
│ │ ├── neuron1.x/
│ │ │ ├── announce-eol-mx-before-1-5.rst
│ │ │ ├── announce-eol-pt-1-5.rst
│ │ │ ├── announce-eol-pt-before-1-8.rst
│ │ │ ├── announce-eol-tf-before-2-5.rst
│ │ │ ├── announce-eol-tf-before-2-7.rst
│ │ │ ├── announcements.rst
│ │ │ ├── eol-ncgs-env_2.rst
│ │ │ ├── eol-pt-15.rst
│ │ │ └── eol-tf-21-24.rst
│ │ └── neuron2.x/
│ │ ├── announce-component-change.rst
│ │ ├── announce-correction-neuron-driver-support-inf1.rst
│ │ ├── announce-deprecation-containers-rtd.rst
│ │ ├── announce-deprecation-nxd-path-trace-api.rst
│ │ ├── announce-deprecation-transformer-flag.rst
│ │ ├── announce-eol-megatron-lm.rst
│ │ ├── announce-eol-python-3-7.rst
│ │ ├── announce-eol-ubuntu-18.rst
│ │ ├── announce-eos-al2.rst
│ │ ├── announce-eos-beta-pytorch-neuroncore-placement-apis.rst
│ │ ├── announce-eos-bf16-vars.rst
│ │ ├── announce-eos-block-dimension-nki.rst
│ │ ├── announce-eos-dlami-ubuntu-22-04.rst
│ │ ├── announce-eos-dlami.rst
│ │ ├── announce-eos-inf1-virtual-environments.rst
│ │ ├── announce-eos-jax-neuronx-nki-call.rst
│ │ ├── announce-eos-megatronlm-2-13.rst
│ │ ├── announce-eos-mllama-checkpoint.rst
│ │ ├── announce-eos-multiframework-dlamis-inf1.rst
│ │ ├── announce-eos-nemo.rst
│ │ ├── announce-eos-neuron-det.rst
│ │ ├── announce-eos-neuron-driver-support-inf1.rst
│ │ ├── announce-eos-neuron-profiler-2.rst
│ │ ├── announce-eos-neuron-profiler-v230.rst
│ │ ├── announce-eos-neuron-profiler.rst
│ │ ├── announce-eos-neurondevice-version.rst
│ │ ├── announce-eos-neurondevice.rst
│ │ ├── announce-eos-nxd-examples.rst
│ │ ├── announce-eos-nxdt-nxd-core-training.rst
│ │ ├── announce-eos-probuf.rst
│ │ ├── announce-eos-pt-versions.rst
│ │ ├── announce-eos-pt2.rst
│ │ ├── announce-eos-python38.rst
│ │ ├── announce-eos-pytorch-1-1-3.rst
│ │ ├── announce-eos-pytorch-1-9.rst
│ │ ├── announce-eos-pytorch-2-1.rst
│ │ ├── announce-eos-pytorch-2-7-2-8-v229.rst
│ │ ├── announce-eos-pytorch-2-7-2-8.rst
│ │ ├── announce-eos-pytorch-profiling-api.rst
│ │ ├── announce-eos-tensorboard-tools.rst
│ │ ├── announce-eos-tensorflow-2-8-9.rst
│ │ ├── announce-eos-tensorflow-inf2.rst
│ │ ├── announce-eos-tensorflow1-x.rst
│ │ ├── announce-eos-torch-neuron.rst
│ │ ├── announce-eos-torch-neuronx-nki-jit.rst
│ │ ├── announce-eos-u20-dlamis.rst
│ │ ├── announce-eos-xla-bf16.rst
│ │ ├── announce-intent-eol-nemo-arg.rst
│ │ ├── announce-intent-eos-opt.rst
│ │ ├── announce-intent-eos-pt-version.rst
│ │ ├── announce-intent-eos-pt2-6.rst
│ │ ├── announce-intent-eos-tensorflow-tutorial-inf.rst
│ │ ├── announce-intent-eos-tnx.rst
│ │ ├── announce-intent-maintenance-tnx.rst
│ │ ├── announce-maintenance-mxnet.rst
│ │ ├── announce-maintenance-nxdi-nxd-core-inference.rst
│ │ ├── announce-maintenance-nxdt-nxd-core-training.rst
│ │ ├── announce-maintenance-tf.rst
│ │ ├── announce-moving-samples.rst
│ │ ├── announce-nki-library-namespace-changes-2-28.rst
│ │ ├── announce-nki-namespace-migration.rst
│ │ ├── announce-no-longer-support-neuron-det.rst
│ │ ├── announce-no-longer-support-nxd-examples.rst
│ │ ├── announce-no-longer-support-pytorch-113.rst
│ │ ├── announce-no-longer-support-pytorch-2-1.rst
│ │ ├── announce-no-longer-support-pytorch-2-7-2-8.rst
│ │ ├── announce-no-longer-support-tensorflow-inf2.rst
│ │ ├── announce-no-longer-support-u20-dlc-dlami.rst
│ │ ├── announce-no-support-al2.rst
│ │ ├── announce-no-support-device-version.rst
│ │ ├── announce-no-support-jax-neuronx-nki-call.rst
│ │ ├── announce-no-support-llama3-2-checkpoint.rst
│ │ ├── announce-no-support-nemo-megatron.rst
│ │ ├── announce-no-support-neurondevice.rst
│ │ ├── announce-no-support-nki-jit-torch.rst
│ │ ├── announce-no-support-tensorboard-plugin.rst
│ │ ├── announce-no-support-tensorflow1-x.rst
│ │ ├── announce-no-support-tensorflow2-10.rst
│ │ ├── announce-no-support-tf-versions.rst
│ │ ├── announce-no-support-torch-neuron-versions.rst
│ │ ├── announce-no-support-ubuntu-20-base.rst
│ │ ├── announce-no-support-vllm-v0.rst
│ │ ├── announce-nxdi-changes.rst
│ │ ├── announce-package-change.rst
│ │ ├── announce-python38-no-longer-support.rst
│ │ ├── announce-transition-pytorch-trainium.rst
│ │ ├── announcement-end-of-support-neuronxcc-nki.rst
│ │ ├── announcement-end-of-support-nxdt-nxd-core.rst
│ │ ├── announcement-end-of-support-parallel-model-trace.rst
│ │ ├── announcement-end-of-support-pytorch-2-6.rst
│ │ ├── announcement-end-of-support-vllm-v0.rst
│ │ ├── announcement-nki-library-kernel-migration.rst
│ │ ├── announcement-nki-library-namespace-changes.rst
│ │ ├── announcement-python-3-9-eol.rst
│ │ ├── dlami-neuron-2.10.rst
│ │ ├── dlami-neuron-2.12.rst
│ │ ├── dlami-pytorch-introduce.rst
│ │ ├── end-of-support-pt2.rst
│ │ ├── github-changes.rst
│ │ ├── gpg-expiration.rst
│ │ ├── neuron-rtd-eol.rst
│ │ ├── neuron2-intro.rst
│ │ ├── neuron230-packages-changes.rst
│ │ ├── neuron250-packages-changes.rst
│ │ ├── release-neuron2.4.rst
│ │ ├── sm-training-dlc-2.9.1.rst
│ │ └── sm-training-trn1-introduce.rst
│ ├── appnotes/
│ │ ├── index.rst
│ │ ├── mxnet-neuron/
│ │ │ └── flex-eg.rst
│ │ ├── neuron-cc/
│ │ │ └── mixed-precision.rst
│ │ ├── neuron1x/
│ │ │ ├── important-neuronx-dkms.txt
│ │ │ └── introducing-libnrt.rst
│ │ ├── neuronx-cc/
│ │ │ └── neuronx-cc-training-mixed-precision.rst
│ │ ├── neuronx-distributed/
│ │ │ ├── introducing-nxd-inference.rst
│ │ │ └── introducing-nxdt-training.rst
│ │ ├── perf/
│ │ │ └── neuron-cc/
│ │ │ ├── parallel-ncgs.rst
│ │ │ └── performance-tuning.rst
│ │ ├── torch-neuron/
│ │ │ ├── bucketing-app-note.rst
│ │ │ ├── index.rst
│ │ │ ├── rcnn-app-note.rst
│ │ │ └── torch-neuron-dataparallel-app-note.rst
│ │ ├── torch-neuronx/
│ │ │ ├── index.rst
│ │ │ ├── introducing-pytorch-2-6.rst
│ │ │ ├── introducing-pytorch-2-7.rst
│ │ │ ├── introducing-pytorch-2-8.rst
│ │ │ ├── introducing-pytorch-2-9.rst
│ │ │ ├── introducing-pytorch-2-x.rst
│ │ │ ├── migration-from-xla-downcast-bf16.rst
│ │ │ ├── torch-neuronx-dataparallel-app-note.rst
│ │ │ └── torch-neuronx-graph-partitioner-app-note.rst
│ │ └── transformers-neuronx/
│ │ └── generative-llm-inference-with-neuron.rst
│ ├── arch/
│ │ ├── glossary.rst
│ │ ├── index.rst
│ │ ├── neuron-features/
│ │ │ ├── custom-c++-operators.rst
│ │ │ ├── data-types.rst
│ │ │ ├── index.rst
│ │ │ ├── logical-neuroncore-config.rst
│ │ │ ├── neuron-caching.rst
│ │ │ ├── neuroncore-batching.rst
│ │ │ ├── neuroncore-pipeline.rst
│ │ │ └── rounding-modes.rst
│ │ └── neuron-hardware/
│ │ ├── inf1-arch.rst
│ │ ├── inf2-arch.rst
│ │ ├── inferentia.rst
│ │ ├── inferentia2.rst
│ │ ├── neuron-core-v1.rst
│ │ ├── neuron-core-v2.rst
│ │ ├── neuron-core-v3.rst
│ │ ├── neuron-core-v4.rst
│ │ ├── trainium.rst
│ │ ├── trainium2.rst
│ │ ├── trainium3.rst
│ │ ├── trn1-arch.rst
│ │ ├── trn2-arch.rst
│ │ └── trn3-arch.rst
│ ├── benchmarks/
│ │ ├── index.rst
│ │ ├── inf1/
│ │ │ ├── data.csv
│ │ │ ├── index.rst
│ │ │ ├── instance_prices.csv
│ │ │ ├── latency_data_encoder.csv
│ │ │ ├── throughput_data_cnn.csv
│ │ │ └── throughput_data_encoder.csv
│ │ ├── inf2/
│ │ │ ├── inf2-performance.rst
│ │ │ ├── inf2_instance_prices.csv
│ │ │ ├── latency_data_decoder.csv
│ │ │ ├── latency_data_encoder.csv
│ │ │ ├── latency_data_encoder_decoder.csv
│ │ │ ├── latency_data_vision.csv
│ │ │ ├── latency_data_vision_cnn.csv
│ │ │ ├── latency_data_vision_dit.csv
│ │ │ ├── latency_data_vision_sd.csv
│ │ │ ├── latency_data_vision_transformers.csv
│ │ │ ├── throughput_data_decoder.csv
│ │ │ ├── throughput_data_encoder.csv
│ │ │ ├── throughput_data_encoder_decoder.csv
│ │ │ ├── throughput_data_vision.csv
│ │ │ ├── throughput_data_vision_cnn.csv
│ │ │ ├── throughput_data_vision_dit.csv
│ │ │ ├── throughput_data_vision_sd.csv
│ │ │ └── throughput_data_vision_transformers.csv
│ │ └── trn1/
│ │ ├── latency_data_decoder.csv
│ │ ├── latency_data_encoder.csv
│ │ ├── latency_data_encoder_decoder.csv
│ │ ├── throughput_data_decoder.csv
│ │ ├── throughput_data_encoder.csv
│ │ ├── throughput_data_encoder_decoder.csv
│ │ ├── training_data_decoder.csv
│ │ ├── training_data_encoder.csv
│ │ ├── training_data_vision_transformers.csv
│ │ ├── trn1-inference-performance.rst
│ │ ├── trn1-training-performance.rst
│ │ ├── trn1_instance_prices.csv
│ │ └── trn1_trn1n_nlp_data.csv
│ ├── beta-participation.rst
│ ├── calculator/
│ │ └── neuron-calculator.rst
│ ├── faq/
│ │ ├── contributing-faq.rst
│ │ ├── index.rst
│ │ ├── inference/
│ │ │ ├── neuron-faq.rst
│ │ │ └── trouble-shooting-faq.rst
│ │ ├── neuron2-intro-faq.rst
│ │ ├── onnx-faq.rst
│ │ ├── roadmap-faq.rst
│ │ └── training/
│ │ └── neuron-training.rst
│ ├── faq.rst
│ ├── index.rst
│ ├── models/
│ │ ├── index.rst
│ │ ├── inference-inf1-samples.rst
│ │ ├── inference-inf2-trn1-samples.rst
│ │ └── training-trn1-samples.rst
│ ├── monitoring-tools.rst
│ ├── news-and-blogs/
│ │ ├── CONTRIBUTING.md
│ │ ├── JIRA-INTEGRATION-DESIGN.md
│ │ ├── README.md
│ │ ├── article-template.yaml
│ │ ├── index.rst
│ │ ├── news-and-blogs.yaml
│ │ └── validate_articles.py
│ ├── oss/
│ │ └── index.rst
│ ├── profiling-tools.rst
│ ├── quick-start/
│ │ ├── _specs/
│ │ │ └── REFACTORING_NOTES.md
│ │ ├── docs-quicklinks.rst
│ │ ├── github-samples.rst
│ │ ├── index.rst
│ │ ├── inference-quickstart.rst
│ │ ├── mxnet-neuron.rst
│ │ ├── tab-inference-tensorflow-neuron.rst
│ │ ├── tensorflow-neuron.rst
│ │ ├── torch-neuron-tab-training.rst
│ │ ├── torch-neuron.rst
│ │ ├── training-quickstart.rst
│ │ └── user-guide-quickstart.rst
│ ├── sdk-policy.rst
│ ├── security.rst
│ ├── troubleshooting.rst
│ ├── what-is-neuron.rst
│ └── whats-new.rst
├── archive/
│ ├── helper-tools/
│ │ ├── index.rst
│ │ ├── tutorial-neuron-check-model.rst
│ │ └── tutorial-neuron-gatherinfo.rst
│ ├── index.rst
│ ├── mxnet-neuron/
│ │ ├── api-compilation-python-api.rst
│ │ ├── api-reference-guide.rst
│ │ ├── api-reference-guide.txt
│ │ ├── developer-guide.rst
│ │ ├── developer-guide.txt
│ │ ├── ec2-then-ec2-devflow.rst
│ │ ├── index.rst
│ │ ├── inference-mxnet-neuron.rst
│ │ ├── inference-mxnet-neuron.txt
│ │ ├── misc-mxnet-neuron.rst
│ │ ├── misc-mxnet-neuron.txt
│ │ ├── mxnet-neuron-setup.rst
│ │ ├── mxnet-neuron-setup.txt
│ │ ├── neo-then-hosting-devflow.rst
│ │ ├── setup/
│ │ │ ├── mxnet-install-prev-al2.rst
│ │ │ ├── mxnet-install-prev-al2023.rst
│ │ │ ├── mxnet-install-prev-u20.rst
│ │ │ ├── mxnet-install-prev-u22.rst
│ │ │ ├── mxnet-install.rst
│ │ │ ├── mxnet-neuron-al2-base-dlami.rst
│ │ │ ├── mxnet-neuron-al2.rst
│ │ │ ├── mxnet-neuron-al2023.rst
│ │ │ ├── mxnet-neuron-ubuntu20-base-dlami.rst
│ │ │ ├── mxnet-neuron-ubuntu20.rst
│ │ │ ├── mxnet-neuron-ubuntu22.rst
│ │ │ ├── mxnet-update-u20.rst
│ │ │ ├── mxnet-update.rst
│ │ │ ├── prev-releases/
│ │ │ │ ├── neuron-1.14.2-mxnet-install.rst
│ │ │ │ ├── neuron-1.15.0-mxnet-install.rst
│ │ │ │ ├── neuron-1.15.1-mxnet-install.rst
│ │ │ │ ├── neuron-1.15.2-mxnet-install.rst
│ │ │ │ ├── neuron-1.16.3-mxnet-install.rst
│ │ │ │ ├── neuron-1.17.2-mxnet-install.rst
│ │ │ │ ├── neuron-1.18.0-mxnet-install.rst
│ │ │ │ └── neuron-1.19.0-mxnet-install.rst
│ │ │ └── setup-inference
│ │ ├── troubleshooting-guide.rst
│ │ └── tutorials/
│ │ ├── mxnet-tutorial-setup.rst
│ │ ├── tutorial-model-serving.rst
│ │ ├── tutorials-mxnet-computervision.rst
│ │ ├── tutorials-mxnet-neuron.rst
│ │ ├── tutorials-mxnet-neuron.txt
│ │ ├── tutorials-mxnet-nlp.rst
│ │ └── tutorials-mxnet-utilizing-neuron-capabilities.rst
│ ├── neuronperf/
│ │ ├── index.rst
│ │ ├── neuronperf_api.rst
│ │ ├── neuronperf_benchmark_guide.rst
│ │ ├── neuronperf_compile_guide.rst
│ │ ├── neuronperf_evaluate_guide.rst
│ │ ├── neuronperf_examples.rst
│ │ ├── neuronperf_faq.rst
│ │ ├── neuronperf_framework_notes.rst
│ │ ├── neuronperf_install.rst
│ │ ├── neuronperf_model_index_guide.rst
│ │ ├── neuronperf_overview.rst
│ │ ├── neuronperf_terminology.rst
│ │ ├── neuronperf_troubleshooting.rst
│ │ ├── rn.rst
│ │ ├── setup.cfg
│ │ ├── setup.py
│ │ ├── test_resnet50_pt.py
│ │ └── test_simple_pt.py
│ ├── src/
│ │ └── benchmark/
│ │ └── pytorch/
│ │ ├── bert-base-cased_benchmark.py
│ │ ├── bert-base-cased_compile.py
│ │ ├── bert-base-uncased_benchmark.py
│ │ ├── bert-base-uncased_compile.py
│ │ ├── distilbert-base-uncased-finetuned-sst-2-english_benchmark.py
│ │ ├── distilbert-base-uncased-finetuned-sst-2-english_compile.py
│ │ ├── distilbert-base-uncased_benchmark.py
│ │ ├── distilbert-base-uncased_compile.py
│ │ ├── distilroberta-base_benchmark.py
│ │ ├── distilroberta-base_compile.py
│ │ ├── hf-google-vit_benchmark.py
│ │ ├── hf-openai-clip_benchmark.py
│ │ ├── hf_pretrained_wav2vec2_conformer_relpos_benchmark.py
│ │ ├── hf_pretrained_wav2vec2_conformer_rope_benchmark.py
│ │ ├── inf2_benchmark.py
│ │ ├── opt_benchmark.py
│ │ ├── perceiver-multimodal_benchmark.py
│ │ ├── perceiver-multimodal_compile.py
│ │ ├── perceiver-vision_benchmark.py
│ │ ├── perceiver-vision_compile.py
│ │ ├── pixart_alpha_benchmark.py
│ │ ├── pixart_sigma_benchmark.py
│ │ ├── resnet50_benchmark.py
│ │ ├── resnet50_compile.py
│ │ ├── resnet_benchmark.py
│ │ ├── resnet_compile.py
│ │ ├── sd2_512_benchmark.py
│ │ ├── sd2_512_compile.py
│ │ ├── sd2_768_benchmark.py
│ │ ├── sd2_768_compile.py
│ │ ├── sd2_inpainting_benchmark.py
│ │ ├── sd2_inpainting_inference.py
│ │ ├── sd_15_512_benchmark.py
│ │ ├── sd_15_512_compile.py
│ │ ├── sd_4x_upscaler_benchmark.py
│ │ ├── sd_4x_upscaler_compile.py
│ │ ├── sdxl_base_1024_benchmark.py
│ │ ├── sdxl_base_1024_compile.py
│ │ ├── sdxl_base_and_refiner_1024_benchmark.py
│ │ ├── sdxl_base_and_refiner_1024_compile.py
│ │ ├── unet_benchmark.py
│ │ ├── unet_compile.py
│ │ ├── vgg_benchmark.py
│ │ └── vgg_compile.py
│ ├── tensorboard/
│ │ └── getting-started-tensorboard-neuron-plugin.rst
│ ├── tensorflow/
│ │ ├── index.rst
│ │ ├── setup-legacy-inf1-tensorflow.rst
│ │ ├── tensorflow-neuron/
│ │ │ ├── additional-examples.rst
│ │ │ ├── additional-examples.txt
│ │ │ ├── api-auto-replication-api.rst
│ │ │ ├── api-compilation-python-api.rst
│ │ │ ├── api-reference-guide.rst
│ │ │ ├── api-reference-guide.txt
│ │ │ ├── api-tfn-analyze-model-api.rst
│ │ │ ├── api-tracing-python-api.rst
│ │ │ ├── dlc-then-ec2-devflow.rst
│ │ │ ├── dlc-then-ecs-devflow.rst
│ │ │ ├── dlc-then-eks-devflow.rst
│ │ │ ├── ec2-then-ec2-devflow.rst
│ │ │ ├── misc-tensorflow-neuron.rst
│ │ │ ├── misc-tensorflow-neuron.txt
│ │ │ ├── neo-then-hosting-devflow.rst
│ │ │ ├── setup/
│ │ │ │ ├── prev-releases/
│ │ │ │ │ ├── neuron-1.14.2-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.15.0-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.15.1-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.15.2-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.16.3-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.17.0-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.17.1-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.17.2-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.18.0-tensorflow-install.rst
│ │ │ │ │ └── neuron-1.19.0-tensorflow-install.rst
│ │ │ │ ├── tensorflow-install-prev-al2023.rst
│ │ │ │ ├── tensorflow-install-prev-u20.rst
│ │ │ │ ├── tensorflow-install-prev-u22.rst
│ │ │ │ ├── tensorflow-install-prev.rst
│ │ │ │ ├── tensorflow-install.rst
│ │ │ │ ├── tensorflow-update-u20.rst
│ │ │ │ ├── tensorflow-update-u22.rst
│ │ │ │ └── tensorflow-update.rst
│ │ │ ├── tensorflow2-accelerated-ops.rst
│ │ │ ├── tf2_faq.rst
│ │ │ └── tutorials/
│ │ │ ├── bert_demo/
│ │ │ │ ├── bert_demo.rst
│ │ │ │ ├── glue_mrpc_dev.tsv
│ │ │ │ └── mrpc.proto
│ │ │ ├── index.rst
│ │ │ ├── k8s_bert_demo/
│ │ │ │ └── Dockerfile.tfserving_example
│ │ │ ├── tensorflow-tutorial-setup.rst
│ │ │ ├── tutorials-tensorflow-neuron.rst
│ │ │ ├── tutorials-tensorflow-neuron.txt
│ │ │ ├── tutorials-tensorflow-nlp.rst
│ │ │ └── tutorials-tensorflow-utilizing-neuron-capabilities.rst
│ │ ├── tensorflow-neuron-inference.rst
│ │ ├── tensorflow-neuron-inference.txt
│ │ ├── tensorflow-neuronx/
│ │ │ ├── api-reference-guide.rst
│ │ │ ├── api-reference-guide.txt
│ │ │ ├── misc-tensorflow-neuronx.rst
│ │ │ ├── misc-tensorflow-neuronx.txt
│ │ │ ├── setup/
│ │ │ │ ├── index.rst
│ │ │ │ ├── prev-releases/
│ │ │ │ │ ├── neuronx-2.8.0-tensorflow-install.rst
│ │ │ │ │ └── neuronx-2.9.0-tensorflow-install.rst
│ │ │ │ ├── tensorflow-install-prev-al2.rst
│ │ │ │ ├── tensorflow-install-prev-al2023.rst
│ │ │ │ ├── tensorflow-install-prev-u20.rst
│ │ │ │ ├── tensorflow-install-prev-u22.rst
│ │ │ │ ├── tensorflow-neuronx-install.rst
│ │ │ │ ├── tensorflow-update-al2-dlami.rst
│ │ │ │ ├── tensorflow-update-al2.rst
│ │ │ │ ├── tensorflow-update-u20-dlami.rst
│ │ │ │ ├── tensorflow-update-u20.rst
│ │ │ │ └── tensorflow-update-u22.rst
│ │ │ ├── tf-neuronx-auto-replication-api.rst
│ │ │ ├── tfneuronx-python-tracing-api.rst
│ │ │ ├── tfnx-analyze-model-api.rst
│ │ │ └── tutorials/
│ │ │ ├── tutorial-tensorflowx-serving-NeuronRT-Visible-Cores.rst
│ │ │ ├── tutorials-tensorflow-neuronx.rst
│ │ │ └── tutorials-tensorflow-neuronx.txt
│ │ ├── tensorflow-neuronx-inference.rst
│ │ ├── tensorflow-neuronx-inference.txt
│ │ ├── tensorflow-setup.rst
│ │ └── tensorflow-setup.txt
│ ├── torch-neuron/
│ │ ├── additional-examples-inference-torch-neuron.rst
│ │ ├── additional-examples-inference-torch-neuron.txt
│ │ ├── api-compilation-python-api.rst
│ │ ├── api-core-placement.rst
│ │ ├── api-reference-guide-torch-neuron.rst
│ │ ├── api-reference-guide-torch-neuron.txt
│ │ ├── api-torch-neuron-dataparallel-api.rst
│ │ ├── developer-guide-torch-neuron.rst
│ │ ├── developer-guide-torch-neuron.txt
│ │ ├── guides/
│ │ │ ├── core-placement/
│ │ │ │ └── torch-core-placement.rst
│ │ │ └── torch-lstm-support.rst
│ │ ├── index.rst
│ │ ├── inference-torch-neuron.rst
│ │ ├── misc-inference-torch-neuron.rst
│ │ ├── misc-inference-torch-neuron.txt
│ │ ├── placement.py
│ │ ├── setup/
│ │ │ ├── index.rst
│ │ │ ├── prev-releases/
│ │ │ │ ├── neuron-1.14.2-pytorch-install.rst
│ │ │ │ ├── neuron-1.15.0-pytorch-install.rst
│ │ │ │ ├── neuron-1.15.1-pytorch-install.rst
│ │ │ │ ├── neuron-1.15.2-pytorch-install.rst
│ │ │ │ ├── neuron-1.16.1-pytorch-install.rst
│ │ │ │ ├── neuron-1.16.2-pytorch-install.rst
│ │ │ │ ├── neuron-1.16.3-pytorch-install.rst
│ │ │ │ ├── neuron-1.17.2-pytorch-install.rst
│ │ │ │ ├── neuron-1.18.0-pytorch-install.rst
│ │ │ │ ├── neuron-1.19.0-pytorch-install.rst
│ │ │ │ ├── neuron-2.3.0-pytorch-install.rst
│ │ │ │ ├── neuron-2.4.0-pytorch-install.rst
│ │ │ │ └── neuron-2.5.0-pytorch-install.rst
│ │ │ ├── pytorch-install-cxx11.rst
│ │ │ ├── pytorch-install-prev-al2.rst
│ │ │ ├── pytorch-install-prev-al2023.rst
│ │ │ ├── pytorch-install-prev-u20.rst
│ │ │ ├── pytorch-install-prev-u22.rst
│ │ │ ├── pytorch-install-prev.rst
│ │ │ ├── pytorch-install.rst
│ │ │ ├── pytorch-update-al2-dlami.rst
│ │ │ ├── pytorch-update-al2023.rst
│ │ │ ├── pytorch-update-u20-dlami.rst
│ │ │ ├── pytorch-update-u20.rst
│ │ │ ├── pytorch-update-u22.rst
│ │ │ └── pytorch-update.rst
│ │ ├── torch-neuron-dataparallel-example-default.rst
│ │ ├── torch-neuron-dataparallel-example-dim-neq-zero.rst
│ │ ├── torch-neuron-dataparallel-example-disable-dynamic-batching.rst
│ │ ├── torch-neuron-dataparallel-example-dynamic-batching.rst
│ │ ├── torch-neuron-dataparallel-example-specify-ncs.rst
│ │ ├── troubleshooting-guide.rst
│ │ └── tutorials/
│ │ ├── neuroncore_pipeline_pytorch.rst
│ │ ├── pytorch-tutorial-setup.rst
│ │ ├── transformers-marianmt.rst
│ │ ├── tutorial-libtorch.rst
│ │ ├── tutorial-torchserve.rst
│ │ ├── tutorial_source_instructions/
│ │ │ ├── run_libtorch.sh
│ │ │ └── run_torchserve_u20.sh
│ │ ├── tutorials-inference-torch-neuron.rst
│ │ ├── tutorials-inference-torch-neuron.txt
│ │ ├── tutorials-torch-neuron-computervision.rst
│ │ ├── tutorials-torch-neuron-nlp.rst
│ │ └── tutorials-utilizing-neuron-capabilities.rst
│ ├── transformers-neuronx/
│ │ ├── api-reference-guide.rst
│ │ ├── api-reference-guide.txt
│ │ ├── developer-guide.rst
│ │ ├── developer-guide.txt
│ │ ├── index.rst
│ │ ├── setup/
│ │ │ └── index.rst
│ │ ├── transformers-neuronx-api-reference.rst
│ │ ├── transformers-neuronx-developer-guide-for-continuous-batching.rst
│ │ ├── transformers-neuronx-developer-guide.rst
│ │ ├── transformers-neuronx-misc.rst
│ │ ├── transformers-neuronx-misc.txt
│ │ ├── transformers-neuronx-tutorials.rst
│ │ ├── transformers-neuronx-tutorials.txt
│ │ └── transformers-neuronx.txt
│ └── tutorials/
│ ├── finetune_t5.rst
│ ├── finetuning_llama2_7b_ptl.rst
│ ├── gpt3_neuronx_nemo_megatron_pretraining.rst
│ ├── megatron_gpt_pretraining.rst
│ ├── multinode-training-model-profiling.rst
│ ├── nxd-source-code/
│ │ ├── gpt_neox_tp_zero1/
│ │ │ ├── gpt_neox_20b.sh
│ │ │ └── gpt_neox_6_9b.sh
│ │ └── llama_tp_pp_ptl/
│ │ ├── llama_2_13b.sh
│ │ ├── llama_2_70b.sh
│ │ ├── llama_2_7b.sh
│ │ └── llama_tp_pp_ptl_setup.sh
│ ├── ssd300_demo/
│ │ ├── requirements.txt
│ │ ├── ssd300_demo.rst
│ │ ├── ssd300_detection.py
│ │ ├── ssd300_evaluation.py
│ │ ├── ssd300_evaluation_client.py
│ │ └── ssd300_model.py
│ ├── training-gpt-neox-20b.rst
│ ├── training-gpt-neox.rst
│ ├── training_codegen25_7b.rst
│ ├── training_llama2_tp_pp_ptl.rst
│ └── tutorial_source_code/
│ └── t5_finetuning/
│ ├── t5_finetuning_32_worker_training_code.sh
│ ├── t5_finetuning_multi_worker_training_code.sh
│ ├── t5_finetuning_setup_code.sh
│ ├── t5_finetuning_single_worker_training_code.sh
│ └── t5_modify_run_summarization_code.sh
├── audit-report.md
├── build.sh
├── compiler/
│ ├── error-codes/
│ │ ├── EARG001.rst
│ │ ├── EBIR023.rst
│ │ ├── EBVF030.rst
│ │ ├── EHCA005.rst
│ │ ├── EOOM001.rst
│ │ ├── EOOM002.rst
│ │ ├── ESFH002.rst
│ │ ├── ESPP004.rst
│ │ ├── ESPP047.rst
│ │ ├── EUOC002.rst
│ │ ├── EVRF001.rst
│ │ ├── EVRF004.rst
│ │ ├── EVRF005.rst
│ │ ├── EVRF006.rst
│ │ ├── EVRF007.rst
│ │ ├── EVRF009.rst
│ │ ├── EVRF010.rst
│ │ ├── EVRF011.rst
│ │ ├── EVRF013.rst
│ │ ├── EVRF015.rst
│ │ ├── EVRF016.rst
│ │ ├── EVRF017.rst
│ │ ├── EVRF018.rst
│ │ ├── EVRF019.rst
│ │ ├── EVRF022.rst
│ │ ├── EVRF031.rst
│ │ ├── EXSP001.rst
│ │ ├── EXTP004.rst
│ │ └── index.rst
│ ├── index.rst
│ ├── neuron-cc/
│ │ ├── api-reference-guide.rst
│ │ ├── command-line-reference.rst
│ │ ├── developer-guide.rst
│ │ └── faq.rst
│ ├── neuron-cc.rst
│ ├── neuronx-cc/
│ │ ├── api-reference-guide/
│ │ │ └── index.rst
│ │ ├── developer-guide.rst
│ │ ├── faq.rst
│ │ └── how-to-convolution-in-unet.rst
│ └── neuronx-cc.rst
├── conf.py
├── containers/
│ ├── container-deployment-flows.rst
│ ├── container-sm-hosting-devflow.rst
│ ├── developerflows.rst
│ ├── developerflows.txt
│ ├── dlc-then-customize-devflow.rst
│ ├── dlc-then-ec2-devflow.rst
│ ├── dlc-then-ecs-devflow.rst
│ ├── dlc-then-eks-devflow.rst
│ ├── dlc-then-k8s-devflow.rst
│ ├── docker-example/
│ │ ├── Dockerfile.device-plugin
│ │ ├── index.rst
│ │ ├── inference/
│ │ │ ├── Dockerfile-inference
│ │ │ ├── Dockerfile-inference-dlc
│ │ │ ├── Dockerfile-inference-dlc.rst
│ │ │ ├── Dockerfile-libmode
│ │ │ ├── Dockerfile-libmode.rst
│ │ │ ├── Dockerfile-tf-serving.rst
│ │ │ ├── Dockerfile.mxnet-serving
│ │ │ ├── Dockerfile.tf-serving
│ │ │ ├── config-properties.rst
│ │ │ ├── config.properties
│ │ │ ├── dockerd-libmode-entrypoint.rst
│ │ │ ├── dockerd-libmode-entrypoint.sh
│ │ │ ├── torchserve-neuron.rst
│ │ │ └── torchserve-neuron.sh
│ │ ├── training/
│ │ │ ├── Dockerfile-training-dlc
│ │ │ ├── Dockerfile-trainium-dlc.rst
│ │ │ ├── mlp.rst
│ │ │ ├── mlp_train.py
│ │ │ └── model.py
│ │ └── v1/
│ │ └── inference/
│ │ ├── Dockerfile-app-rt-diff.rst
│ │ ├── Dockerfile-app-rt-same.rst
│ │ ├── Dockerfile-neuron-rtd.rst
│ │ ├── Dockerfile-torch-neuron.rst
│ │ ├── Dockerfile.app-rt-diff
│ │ ├── Dockerfile.neuron-rtd
│ │ ├── Dockerfile.torch-neuron
│ │ ├── dockerd-entrypoint-app-rt-same.rst
│ │ └── dockerd-entrypoint.sh
│ ├── ec2-then-ec2-devflow.rst
│ ├── ec2.rst
│ ├── faq-troubleshooting-releasenote.rst
│ ├── faq.rst
│ ├── files/
│ │ ├── index-dra.rst
│ │ ├── manifests/
│ │ │ ├── clusterrole.yaml
│ │ │ ├── clusterrolebinding.yaml
│ │ │ ├── daemonset.yaml
│ │ │ ├── deviceclass.yaml
│ │ │ ├── namespace.yaml
│ │ │ └── serviceaccount.yaml
│ │ ├── scripts/
│ │ │ └── install-dra-driver.sh
│ │ └── specs/
│ │ ├── 1x4-connected-devices.yaml
│ │ ├── 2-node-inference-us.yaml
│ │ ├── 4-node-inference-us.yaml
│ │ ├── all-devices.yaml
│ │ ├── lnc-setting-trn2.yaml
│ │ ├── specific-driver-version.yaml
│ │ └── us-and-lnc-config.yaml
│ ├── get-started/
│ │ ├── quickstart-configure-deploy-dlc.rst
│ │ └── quickstart-pytorch-inference-dlc.rst
│ ├── getting-started.rst
│ ├── how-to/
│ │ └── how-to-ultraserver.rst
│ ├── index.rst
│ ├── k8.rst
│ ├── kubernetes-getting-started.rst
│ ├── locate-neuron-dlc-image.rst
│ ├── neo-then-hosting-devflow.rst
│ ├── neuron-dra.rst
│ ├── neuron-plugins.rst
│ ├── neuron_dlc_images.csv
│ ├── troubleshooting.rst
│ ├── tutorial-docker-runtime1.0.rst
│ ├── tutorials/
│ │ ├── build-run-neuron-container.rst
│ │ ├── inference/
│ │ │ ├── index.rst
│ │ │ ├── index.txt
│ │ │ ├── k8s_rn50_demo.rst
│ │ │ └── tutorial-infer.rst
│ │ ├── k8s-default-scheduler.rst
│ │ ├── k8s-multiple-scheduler.rst
│ │ ├── k8s-neuron-device-plugin.rst
│ │ ├── k8s-neuron-helm-chart.rst
│ │ ├── k8s-neuron-monitor.rst
│ │ ├── k8s-neuron-problem-detector-and-recovery-irsa.rst
│ │ ├── k8s-neuron-problem-detector-and-recovery.rst
│ │ ├── k8s-neuron-scheduler-flow.rst
│ │ ├── k8s-neuron-scheduler.rst
│ │ ├── k8s-prerequisite.rst
│ │ ├── k8s-setup.rst
│ │ ├── training/
│ │ │ ├── index.rst
│ │ │ ├── index.txt
│ │ │ ├── k8s_mlp_train_demo.rst
│ │ │ └── tutorial-training.rst
│ │ ├── tutorial-docker-env-setup.rst
│ │ └── tutorial-oci-hook.rst
│ └── tutorials.rst
├── devflows/
│ ├── aws-batch-flows.rst
│ ├── aws-batch-flows.txt
│ ├── dlc-then-customize-devflow.rst
│ ├── ec2-flows.rst
│ ├── ec2-flows.txt
│ ├── ecs-flows.rst
│ ├── eks-flows.rst
│ ├── index.rst
│ ├── inference/
│ │ ├── aws-batch-flows.rst
│ │ ├── aws-batch-flows.txt
│ │ ├── byoc-hosting-devflow-inf2.rst
│ │ ├── byoc-hosting-devflow.rst
│ │ ├── container-sm-hosting-devflow.rst
│ │ ├── dev-flows.rst
│ │ ├── dlc-then-ec2-devflow.rst
│ │ ├── dlc-then-ecs-devflow.rst
│ │ ├── dlc-then-eks-devflow.rst
│ │ ├── dlc-then-k8s-devflow.rst
│ │ ├── ec2-flows.rst
│ │ ├── ec2-flows.txt
│ │ ├── ec2-then-ec2-devflow-inf2.rst
│ │ ├── ec2-then-ec2-devflow.rst
│ │ ├── env-setup-text.rst
│ │ ├── neo-then-hosting-devflow.rst
│ │ ├── parallelcluster-flows.rst
│ │ ├── parallelcluster-flows.txt
│ │ ├── sagemaker-flows.rst
│ │ └── sagemaker-flows.txt
│ ├── parallelcluster-flows.rst
│ ├── parallelcluster-flows.txt
│ ├── plugins/
│ │ ├── npd-ecs-flows.rst
│ │ └── npd-ecs-flows.txt
│ ├── sagemaker-flows.rst
│ ├── setup/
│ │ ├── ecs-flows.rst
│ │ ├── ecs-flows.txt
│ │ ├── eks-flows.rst
│ │ └── eks-flows.txt
│ ├── third-party-solutions.rst
│ └── training/
│ ├── aws-batch-flows.rst
│ ├── aws-batch-flows.txt
│ ├── batch/
│ │ └── batch-training.rst
│ ├── dlc-then-ecs-devflow.rst
│ ├── ec2/
│ │ └── ec2-training.rst
│ ├── ec2-flows.rst
│ ├── ec2-flows.txt
│ ├── parallelcluster/
│ │ └── parallelcluster-training.rst
│ ├── parallelcluster-flows.rst
│ ├── parallelcluster-flows.txt
│ ├── sagemaker-flows.rst
│ ├── sagemaker-flows.txt
│ └── sm-devflow/
│ └── sm-training-devflow.rst
├── dlami/
│ └── index.rst
├── frameworks/
│ ├── index.rst
│ ├── jax/
│ │ ├── api-reference-guide/
│ │ │ ├── index.rst
│ │ │ └── neuron-envvars.rst
│ │ ├── index.rst
│ │ └── setup/
│ │ ├── jax-neuronx-known-issues.rst
│ │ └── jax-setup.rst
│ └── torch/
│ ├── about/
│ │ └── index.rst
│ ├── guide-torch-neuron-vs-torch-neuronx-inference.rst
│ ├── index.rst
│ ├── inference-torch-neuronx.rst
│ ├── pytorch-native-overview.rst
│ ├── torch-neuronx/
│ │ ├── additional-examples-inference-torch-neuronx.rst
│ │ ├── additional-examples-training.rst
│ │ ├── api-reference-guide/
│ │ │ ├── inference/
│ │ │ │ ├── api-torch-neuronx-analyze.rst
│ │ │ │ ├── api-torch-neuronx-async-lazy-load.rst
│ │ │ │ ├── api-torch-neuronx-core-placement.rst
│ │ │ │ ├── api-torch-neuronx-data-parallel.rst
│ │ │ │ ├── api-torch-neuronx-replace-weights.rst
│ │ │ │ ├── api-torch-neuronx-trace.rst
│ │ │ │ └── inference-api-guide-torch-neuronx.rst
│ │ │ ├── torch-neuronx-profiling-api.rst
│ │ │ └── training/
│ │ │ ├── index.rst
│ │ │ ├── pytorch-neuron-parallel-compile.rst
│ │ │ └── torch-neuron-envvars.rst
│ │ ├── misc-inference-torch-neuronx.rst
│ │ ├── misc-training.rst
│ │ ├── programming-guide/
│ │ │ ├── inference/
│ │ │ │ ├── autobucketing-dev-guide.rst
│ │ │ │ ├── core-placement.rst
│ │ │ │ ├── index.rst
│ │ │ │ └── trace-vs-xla-lazytensor.rst
│ │ │ ├── torch-neuronx-profiling-dev-guide.rst
│ │ │ └── training/
│ │ │ ├── index.rst
│ │ │ ├── pytorch-neuron-debug.rst
│ │ │ └── pytorch-neuron-programming-guide.rst
│ │ ├── pytorch-neuron-supported-operators.rst
│ │ ├── setup/
│ │ │ ├── install-templates/
│ │ │ │ └── pytorch-dev-install.txt
│ │ │ ├── note-setup-general.rst
│ │ │ ├── prev-releases/
│ │ │ │ ├── neuronx-2.7.0-pytorch-install.rst
│ │ │ │ ├── neuronx-2.8.0-pytorch-install.rst
│ │ │ │ └── neuronx-2.9.0-pytorch-install.rst
│ │ │ ├── pytorch-install-prev-al2.rst
│ │ │ ├── pytorch-install-prev-al2023.rst
│ │ │ ├── pytorch-install-prev-u20.rst
│ │ │ ├── pytorch-install-prev-u22.rst
│ │ │ ├── pytorch-install-prev-u24.rst
│ │ │ ├── pytorch-install.rst
│ │ │ ├── pytorch-neuronx-install-cxx11.rst
│ │ │ ├── pytorch-update-al2-dlami.rst
│ │ │ ├── pytorch-update-al2.rst
│ │ │ ├── pytorch-update-al2023.rst
│ │ │ ├── pytorch-update-u20-dlami.rst
│ │ │ ├── pytorch-update-u20.rst
│ │ │ ├── pytorch-update-u22.rst
│ │ │ └── pytorch-update-u24.rst
│ │ ├── setup-trn1-multi-node-execution.rst
│ │ ├── torch-neuronx-dataparallel-example-default.rst
│ │ ├── torch-neuronx-dataparallel-example-dim-neq-zero.rst
│ │ ├── torch-neuronx-dataparallel-example-disable-dynamic-batching.rst
│ │ ├── torch-neuronx-dataparallel-example-dynamic-batching.rst
│ │ ├── torch-neuronx-dataparallel-example-specify-ncs.rst
│ │ ├── training-troubleshooting.rst
│ │ └── tutorials/
│ │ ├── inference/
│ │ │ ├── tutorial-torchserve-neuronx.rst
│ │ │ └── tutorials-torch-neuronx.rst
│ │ ├── note-performance.txt
│ │ └── training/
│ │ ├── analyze_for_training.rst
│ │ ├── bert.rst
│ │ ├── finetune_hftrainer.rst
│ │ ├── mlp.rst
│ │ ├── tutorial_source_code/
│ │ │ ├── analyze_training/
│ │ │ │ └── analyze_training_code.sh
│ │ │ ├── bert_mrpc_finetuning/
│ │ │ │ ├── bert_mrpc_finetuning_converted_checkpoint_training.sh
│ │ │ │ ├── bert_mrpc_finetuning_multi_worker_training_code.sh
│ │ │ │ ├── bert_mrpc_finetuning_setup_code.sh
│ │ │ │ └── bert_mrpc_finetuning_single_worker_training.sh
│ │ │ ├── bert_training/
│ │ │ │ ├── bert_amp_training_code.sh
│ │ │ │ ├── bert_lamb_bf16_training_code.sh
│ │ │ │ ├── bert_lamb_training_code.sh
│ │ │ │ ├── bert_phase2_training_code.sh
│ │ │ │ ├── bert_precompilation_code.sh
│ │ │ │ ├── bert_setup_code.sh
│ │ │ │ ├── bert_setup_code_ph2.sh
│ │ │ │ └── bert_training_code.sh
│ │ │ ├── multi_layer_perceptron_training/
│ │ │ │ └── multi_layer_perceptron_training_code.sh
│ │ │ └── zero1_training/
│ │ │ └── zero1_single_node_training_code.sh
│ │ ├── tutorials-training-torch-neuronx.rst
│ │ └── zero1_gpt2.rst
│ ├── torch-setup.rst
│ └── training-torch-neuronx.rst
├── general/
│ └── faq.rst
├── includes/
│ └── setup/
│ ├── select-framework-note.txt
│ ├── tab-inference-mxnet-neuron-al2.txt
│ ├── tab-inference-mxnet-neuron-al2023.txt
│ ├── tab-inference-mxnet-neuron-u20.txt
│ ├── tab-inference-mxnet-neuron-u22.txt
│ ├── tab-inference-mxnet-neuron.txt
│ ├── tab-inference-tensorflow-neuron-al2.txt
│ ├── tab-inference-tensorflow-neuron-al2023.txt
│ ├── tab-inference-tensorflow-neuron-u20.txt
│ ├── tab-inference-tensorflow-neuron-u22.txt
│ ├── tab-inference-tensorflow-neuronx-al2.txt
│ ├── tab-inference-tensorflow-neuronx-al2023.txt
│ ├── tab-inference-tensorflow-neuronx-u20.txt
│ ├── tab-inference-tensorflow-neuronx-u22.txt
│ ├── tab-inference-torch-neuron-al2.txt
│ ├── tab-inference-torch-neuron-al2023.txt
│ ├── tab-inference-torch-neuron-u20.txt
│ ├── tab-inference-torch-neuron-u22.txt
│ ├── tab-inference-torch-neuron.txt
│ ├── tab-inference-torch-neuronx-al2.txt
│ ├── tab-inference-torch-neuronx-al2023.txt
│ ├── tab-inference-torch-neuronx-u20.txt
│ ├── tab-inference-torch-neuronx-u22.txt
│ └── tab-inference-torch-neuronx-u24.txt
├── index.rst
├── info/
│ └── exclude
├── libraries/
│ ├── index.rst
│ ├── nemo-megatron/
│ │ └── index.rst
│ ├── neuronx-distributed/
│ │ ├── activation_memory_reduction.rst
│ │ ├── activation_memory_reduction_developer_guide.rst
│ │ ├── api-reference-guide-inference.rst
│ │ ├── api-reference-guide-training.rst
│ │ ├── api-reference-guide.rst
│ │ ├── api-reference-guide.txt
│ │ ├── api_guide.rst
│ │ ├── app_notes.rst
│ │ ├── app_notes.txt
│ │ ├── context_parallelism_overview.rst
│ │ ├── developer-guide-inference.rst
│ │ ├── developer-guide-inference.txt
│ │ ├── developer-guide-training.rst
│ │ ├── developer-guide-training.txt
│ │ ├── developer-guide.rst
│ │ ├── developer-guide.txt
│ │ ├── index-inference.rst
│ │ ├── index-training.rst
│ │ ├── lora_finetune_developer_guide.rst
│ │ ├── model_builder_v2_api_reference.rst
│ │ ├── model_optimizer_wrapper_developer_guide.rst
│ │ ├── neuronx-distributed-misc.rst
│ │ ├── neuronx-distributed-misc.txt
│ │ ├── neuronx_distributed_inference_developer_guide.rst
│ │ ├── pipeline_parallelism_overview.rst
│ │ ├── pp_developer_guide.rst
│ │ ├── ptl_developer_guide.rst
│ │ ├── save_load_developer_guide.rst
│ │ ├── setup/
│ │ │ ├── index.rst
│ │ │ └── index.txt
│ │ ├── standard_mixed_precision.rst
│ │ ├── tensor_parallelism_overview.rst
│ │ ├── tp_developer_guide.rst
│ │ └── tutorials/
│ │ ├── finetune_llama3_8b_ptl_lora.rst
│ │ ├── index.rst
│ │ ├── index.txt
│ │ ├── inference.rst
│ │ ├── inference_tutorials.rst
│ │ ├── neuronx_distributed_tutorials.txt
│ │ ├── nxd-source-code/
│ │ │ ├── llama_tp_pp/
│ │ │ │ ├── llama_2_13b.sh
│ │ │ │ ├── llama_2_70b.sh
│ │ │ │ ├── llama_31_70b.sh
│ │ │ │ ├── llama_3_70b.sh
│ │ │ │ └── llama_tp_pp_setup.sh
│ │ │ └── llama_tp_zero1/
│ │ │ ├── llama_2_7b.sh
│ │ │ ├── llama_31_8b.sh
│ │ │ ├── llama_3_8b.sh
│ │ │ └── llama_tp_zero1_setup.sh
│ │ ├── nxd_inference_tutorials.txt
│ │ ├── nxd_training_tutorials.txt
│ │ ├── training.rst
│ │ ├── training_llama_tp_pp.rst
│ │ ├── training_llama_tp_zero1.rst
│ │ └── training_tutorials.rst
│ ├── nxd-inference/
│ │ ├── _templates/
│ │ │ ├── model_card.jinja.rst
│ │ │ └── model_card_qwen3.jinja.rst
│ │ ├── api-guides/
│ │ │ ├── api-guide.rst
│ │ │ ├── api-guide.txt
│ │ │ └── index.rst
│ │ ├── app-notes/
│ │ │ ├── app_notes.txt
│ │ │ ├── index.rst
│ │ │ └── parallelism.rst
│ │ ├── developer_guides/
│ │ │ ├── accuracy-eval-with-datasets.rst
│ │ │ ├── custom-quantization.rst
│ │ │ ├── disaggregated-inference.rst
│ │ │ ├── feature-guide.rst
│ │ │ ├── how-to-use-fpem.rst
│ │ │ ├── index.rst
│ │ │ ├── llm-inference-benchmarking-guide.rst
│ │ │ ├── migrate-from-tnx-to-nxdi.rst
│ │ │ ├── model-reference.rst
│ │ │ ├── moe-arch-deep-dive.rst
│ │ │ ├── nxd-examples-migration-guide.rst
│ │ │ ├── onboarding-models.rst
│ │ │ ├── performance-cli-params.rst
│ │ │ ├── vllm-user-guide-v1.rst
│ │ │ ├── vllm-user-guide.rst
│ │ │ ├── weights-sharding-guide.rst
│ │ │ └── writing-tests.rst
│ │ ├── examples/
│ │ │ └── vllm_client.py
│ │ ├── index.rst
│ │ ├── misc/
│ │ │ ├── index.rst
│ │ │ ├── misc.txt
│ │ │ └── nxdi-troubleshooting.rst
│ │ ├── models/
│ │ │ ├── index.rst
│ │ │ ├── llama3/
│ │ │ │ ├── data/
│ │ │ │ │ └── card_llama33_70b.yml
│ │ │ │ └── llama_33_70b.rst
│ │ │ ├── models.txt
│ │ │ └── qwen3/
│ │ │ ├── data/
│ │ │ │ └── card_qwen3_moe_235b.yml
│ │ │ └── qwen3_moe_235b.rst
│ │ ├── neuron-inference-overview.rst
│ │ ├── nxdi-setup.rst
│ │ ├── overview-index.rst
│ │ ├── setup.txt
│ │ ├── tutorials/
│ │ │ ├── disaggregated-inference-tutorial-1p1d.rst
│ │ │ ├── disaggregated-inference-tutorial.rst
│ │ │ ├── flux-inference-tutorial.ipynb
│ │ │ ├── flux-inpainting-inference-tutorial.ipynb
│ │ │ ├── generating-results-with-performance-cli.ipynb
│ │ │ ├── index.rst
│ │ │ ├── llama4-tutorial-v0.ipynb
│ │ │ ├── llama4-tutorial.ipynb
│ │ │ ├── llama405b_perf_comparison.csv
│ │ │ ├── llama70b_apc_perf_comparison.csv
│ │ │ ├── llama70b_perf_comparison.csv
│ │ │ ├── modules_to_not_convert.json
│ │ │ ├── pixtral-tutorial.ipynb
│ │ │ ├── qwen2-vl-tutorial.ipynb
│ │ │ ├── qwen3-moe-tutorial.ipynb
│ │ │ ├── qwen3-vl-tutorial.ipynb
│ │ │ ├── sd-inference-tutorial.rst
│ │ │ ├── trn1-llama3.1-70b-instruct-accuracy-eval-tutorial.ipynb
│ │ │ ├── trn2-llama3.1-405b-speculative-tutorial.rst
│ │ │ ├── trn2-llama3.1-405b-tutorial.rst
│ │ │ ├── trn2-llama3.1-8b-multi-lora-tutorial.ipynb
│ │ │ ├── trn2-llama3.3-70b-apc-tutorial.ipynb
│ │ │ ├── trn2-llama3.3-70b-dp-tutorial.ipynb
│ │ │ ├── trn2-llama3.3-70b-fp8.rst
│ │ │ ├── trn2-llama3.3-70b-tutorial.rst
│ │ │ └── trn3-gpt-oss-120b-tutorial.rst
│ │ └── vllm/
│ │ ├── index.rst
│ │ ├── quickstart-vllm-offline-serving.rst
│ │ └── quickstart-vllm-online-serving.rst
│ ├── nxd-training/
│ │ ├── api-guide.txt
│ │ ├── api-reference-guide.rst
│ │ ├── app_notes/
│ │ │ ├── nxd-training-amr-appnote.rst
│ │ │ ├── nxd-training-cp-appnote.rst
│ │ │ ├── nxd-training-pp-appnote.rst
│ │ │ └── nxd-training-tp-appnote.rst
│ │ ├── app_notes.rst
│ │ ├── app_notes.txt
│ │ ├── developer-guide.rst
│ │ ├── developer_guides/
│ │ │ ├── cpu_mode_developer_guide.rst
│ │ │ ├── dev-guide.txt
│ │ │ ├── index.rst
│ │ │ ├── migration_nemo_nxdt.rst
│ │ │ ├── migration_nnm_nxdt.rst
│ │ │ ├── nemo_nxdt_mapping.csv
│ │ │ ├── new_dataloader_guide.rst
│ │ │ ├── new_model_guide.rst
│ │ │ ├── nnm_nxdt_mapping.csv
│ │ │ └── optimizer_lr_scheduler_flow.rst
│ │ ├── general/
│ │ │ ├── config_overview.rst
│ │ │ ├── features.rst
│ │ │ ├── installation_guide.rst
│ │ │ ├── known-issues.txt
│ │ │ └── known_issues.rst
│ │ ├── index.rst
│ │ ├── misc.rst
│ │ ├── misc.txt
│ │ ├── overview.rst
│ │ ├── overview.txt
│ │ ├── setup.txt
│ │ └── tutorials/
│ │ ├── checkpoint_conversion.rst
│ │ ├── hf_llama3_70B_pretraining.rst
│ │ ├── hf_llama3_8B_DPO_ORPO.rst
│ │ ├── hf_llama3_8B_SFT.rst
│ │ ├── hf_llama3_8B_SFT_LORA.rst
│ │ ├── hf_llama3_8B_pretraining.rst
│ │ ├── index.rst
│ │ ├── megatron_gpt_pretraining.rst
│ │ └── tutorials.txt
│ └── transformers-neuronx/
│ └── index.rst
├── llms.txt
├── neuron-customops/
│ ├── api-reference-guide/
│ │ ├── api-reference-guide.rst
│ │ └── custom-ops-ref-guide.rst
│ ├── customops-intro.txt
│ ├── index.rst
│ ├── misc-customops.rst
│ ├── programming-guide/
│ │ ├── custom-c++-operators-devguide.rst
│ │ └── programming-guide.rst
│ └── tutorials/
│ ├── customop-mlp-perf-opt.rst
│ ├── customop-mlp-training.rst
│ ├── tutorial_source_code/
│ │ ├── custom_c_mlp_training/
│ │ │ └── custom_c_mlp_training_code.sh
│ │ └── custom_c_perf_optimization/
│ │ └── custom_c_perf_optimization_code.sh
│ └── tutorials.rst
├── neuron-runtime/
│ ├── about/
│ │ ├── collectives.rst
│ │ ├── core-dump.rst
│ │ └── index.rst
│ ├── api/
│ │ ├── debug-stream-api.rst
│ │ ├── index.rst
│ │ ├── ndebug_stream.rst
│ │ ├── ndl.rst
│ │ ├── nec.rst
│ │ ├── neuron_driver_shared.rst
│ │ ├── neuron_driver_shared_tensor_batch_op.rst
│ │ ├── neuron_ds.rst
│ │ ├── nrt-async-api-best-practices.rst
│ │ ├── nrt-async-api-examples.rst
│ │ ├── nrt-async-api-overview.rst
│ │ ├── nrt.rst
│ │ ├── nrt_async.rst
│ │ ├── nrt_async_sendrecv.rst
│ │ ├── nrt_experimental.rst
│ │ ├── nrt_profile.rst
│ │ ├── nrt_status.rst
│ │ ├── nrt_sys_trace.rst
│ │ └── nrt_version.rst
│ ├── configuration-guide.rst
│ ├── explore/
│ │ ├── compute-comm-overlap.rst
│ │ ├── core-dump-deep-dive.rst
│ │ ├── device-memory.rst
│ │ ├── direct-hbm-tensor-alloc.rst
│ │ ├── index.rst
│ │ ├── internode-collective-comm.rst
│ │ ├── intranode-collective-comm.rst
│ │ ├── runtime-performance-tips.rst
│ │ └── work-with-neff-files.rst
│ ├── faq.rst
│ ├── index.rst
│ ├── nrt-configurable-parameters.rst
│ ├── nrt-developer-guide.rst
│ ├── nrt-troubleshoot.rst
│ └── rn.rst
├── nki/
│ ├── _ext/
│ │ └── nki_directives.py
│ ├── _templates/
│ │ ├── nki-custom-class-attr-only-template.rst
│ │ └── nki-custom-class-template.rst
│ ├── api/
│ │ ├── index.rst
│ │ ├── nki/
│ │ │ ├── __init__.py
│ │ │ ├── collectives/
│ │ │ │ └── __init__.py
│ │ │ ├── isa/
│ │ │ │ └── __init__.py
│ │ │ └── language/
│ │ │ └── __init__.py
│ │ ├── nki.api.shared.rst
│ │ ├── nki.collectives.rst
│ │ ├── nki.isa.rst
│ │ ├── nki.isa.rst.bak
│ │ ├── nki.language.rst
│ │ ├── nki.language.tile_size.rst
│ │ ├── nki.rst
│ │ └── nki.simulate.rst
│ ├── deep-dives/
│ │ ├── index.rst
│ │ ├── mxfp-matmul.rst
│ │ ├── nki-aps.rst
│ │ ├── nki-compiler.rst
│ │ ├── nki-dge.rst
│ │ ├── nki-dma-bandwidth-guide.rst
│ │ ├── nki-dynamic-loops.rst
│ │ ├── nki_perf_guide.rst
│ │ └── src/
│ │ └── mxfp-matmul/
│ │ ├── mx_cpu_utils.py
│ │ ├── mx_kernel_utils.py
│ │ ├── mx_kernels.py
│ │ └── mx_toplevel.py
│ ├── examples/
│ │ ├── average_pool2d/
│ │ │ ├── average_pool2d_jax.py
│ │ │ ├── average_pool2d_nki_kernels.py
│ │ │ └── average_pool2d_torch.py
│ │ ├── fused_mamba/
│ │ │ ├── mamba_nki_kernels.py
│ │ │ └── mamba_torch.py
│ │ ├── getting_started_baremetal.py
│ │ ├── getting_started_jax.py
│ │ ├── getting_started_torch.py
│ │ ├── index-case-1.py
│ │ ├── index-case-3.py
│ │ ├── layout-dynamic-loop.py
│ │ ├── layout-loop.py
│ │ ├── layout-pass.py
│ │ ├── layout-violation.py
│ │ ├── matrix_multiplication/
│ │ │ ├── matrix_multiplication_nki_kernels.py
│ │ │ └── matrix_multiplication_torch.py
│ │ ├── simulate/
│ │ │ └── nki_simulate_example.py
│ │ ├── tensor_addition/
│ │ │ └── tensor_addition_nki_kernels.py
│ │ └── transpose2d/
│ │ ├── transpose2d_jax.py
│ │ ├── transpose2d_nki_kernels.py
│ │ └── transpose2d_torch.py
│ ├── get-started/
│ │ ├── about/
│ │ │ ├── data-representation-overview.rst
│ │ │ ├── index.rst
│ │ │ ├── indexing-overview.rst
│ │ │ ├── lnc.rst
│ │ │ ├── memory-hierarchy-overview.rst
│ │ │ ├── nki-dma-overview.rst
│ │ │ └── tiling-overview.rst
│ │ ├── index.rst
│ │ ├── nki-language-guide.rst
│ │ ├── quickstart-implement-run-kernel.rst
│ │ └── setup-env.rst
│ ├── guides/
│ │ ├── architecture/
│ │ │ ├── index.rst
│ │ │ ├── trainium2_arch.rst
│ │ │ ├── trainium3_arch.rst
│ │ │ └── trainium_inferentia2_arch.rst
│ │ ├── framework_custom_op.rst
│ │ ├── how-to-scheduling-apis.rst
│ │ ├── index.rst
│ │ ├── nki_simulator.rst
│ │ ├── tutorials/
│ │ │ ├── average_pool2d.rst
│ │ │ ├── fused_mamba.rst
│ │ │ ├── index.rst
│ │ │ ├── kernel-optimization.rst
│ │ │ ├── matrix_multiplication.rst
│ │ │ └── transpose2d.rst
│ │ └── use-neuron-profile.rst
│ ├── index.rst
│ ├── library/
│ │ ├── about/
│ │ │ └── index.rst
│ │ ├── api/
│ │ │ ├── attention-block-tkg.rst
│ │ │ ├── attention-cte.rst
│ │ │ ├── attention-tkg.rst
│ │ │ ├── blockwise-mm-backward.rst
│ │ │ ├── conv1d.rst
│ │ │ ├── cross-entropy.rst
│ │ │ ├── cumsum.rst
│ │ │ ├── depthwise-conv1d.rst
│ │ │ ├── dynamic-elementwise-add.rst
│ │ │ ├── fg-allgather.rst
│ │ │ ├── fgcc.rst
│ │ │ ├── find-nonzero-indices.rst
│ │ │ ├── index.rst
│ │ │ ├── mlp.rst
│ │ │ ├── moe-cte.rst
│ │ │ ├── moe-tkg.rst
│ │ │ ├── output-projection-cte.rst
│ │ │ ├── output-projection-tkg.rst
│ │ │ ├── qkv.rst
│ │ │ ├── rmsnorm-quant.rst
│ │ │ ├── rope.rst
│ │ │ ├── router-topk.rst
│ │ │ ├── sb2sb-allgather.rst
│ │ │ ├── topk-reduce.rst
│ │ │ └── transformer-tkg.rst
│ │ ├── index.rst
│ │ ├── kernel-utils/
│ │ │ ├── allocator.rst
│ │ │ ├── index.rst
│ │ │ └── tensor-view.rst
│ │ └── specs/
│ │ ├── design-rmsnorm-quant.rst
│ │ └── index.rst
│ ├── migration/
│ │ ├── index.rst
│ │ ├── nki-0-3-0-update-guide.rst
│ │ ├── nki-beta2-migration-guide.rst
│ │ └── nki_block_dimension_migration_guide.rst
│ ├── nki_faq.rst
│ ├── scripts/
│ │ ├── markdown2rst.py
│ │ └── requirements.txt
│ └── test/
│ ├── test_nki_isa_activation.py
│ ├── test_nki_isa_affine_select.py
│ ├── test_nki_isa_bn_stats.py
│ ├── test_nki_isa_copypredicated.py
│ ├── test_nki_isa_dma_copy.py
│ ├── test_nki_isa_dma_transpose.py
│ ├── test_nki_isa_dropout.py
│ ├── test_nki_isa_iota.py
│ ├── test_nki_isa_local_gather.py
│ ├── test_nki_isa_max8.py
│ ├── test_nki_isa_memset.py
│ ├── test_nki_isa_nc_find_index8.py
│ ├── test_nki_isa_nc_match_replace8.py
│ ├── test_nki_isa_nc_matmul.py
│ ├── test_nki_isa_nc_stream_shuffle.py
│ ├── test_nki_isa_nc_transpose.py
│ ├── test_nki_isa_partition_reduce.py
│ ├── test_nki_isa_range_select.py
│ ├── test_nki_isa_reciprocal.py
│ ├── test_nki_isa_reduce.py
│ ├── test_nki_isa_select_reduce.py
│ ├── test_nki_isa_sequence_bounds.py
│ ├── test_nki_isa_tensor_copy.py
│ ├── test_nki_isa_tensor_scalar.py
│ ├── test_nki_isa_tensor_scalar_cumulative.py
│ ├── test_nki_isa_tensor_tensor.py
│ ├── test_nki_isa_tensor_tensor_scan.py
│ ├── test_nki_mask.py
│ ├── test_nki_memory_semantics.py
│ ├── test_nki_nl_add.py
│ ├── test_nki_nl_atomic_rmw.py
│ ├── test_nki_nl_broadcast.py
│ ├── test_nki_nl_dslice.py
│ ├── test_nki_nl_gather_flattened.py
│ ├── test_nki_nl_load_store.py
│ ├── test_nki_nl_load_store_indirect.py
│ ├── test_nki_nl_load_transpose2d.py
│ ├── test_nki_nl_mgrid.py
│ ├── test_nki_simulate_kernel.py
│ ├── test_nki_spmd_grid.py
│ ├── test_psum_modulo_alloc.py
│ └── test_sbuf_modulo_alloc.py
├── release-notes/
│ ├── 2.29.0.rst
│ ├── archive/
│ │ ├── customcxxps/
│ │ │ ├── gpsimd-customop-lib.rst
│ │ │ └── gpsimd-tools.rst
│ │ ├── index.rst
│ │ ├── libneuronxla.rst
│ │ ├── mxnet-neuron.rst
│ │ ├── nemo/
│ │ │ ├── index.rst
│ │ │ └── neuronx-nemo.rst
│ │ ├── neuron-cc/
│ │ │ ├── neuron-cc-ops/
│ │ │ │ ├── index.rst
│ │ │ │ ├── neuron-cc-ops-mxnet.rst
│ │ │ │ ├── neuron-cc-ops-pytorch.rst
│ │ │ │ ├── neuron-cc-ops-tensorflow.rst
│ │ │ │ └── neuron-cc-ops-xla.rst
│ │ │ └── neuron-cc.rst
│ │ ├── neuron1/
│ │ │ ├── _legacy-labels.rst
│ │ │ ├── neuronrelease/
│ │ │ │ └── previous-content.rst
│ │ │ └── prev/
│ │ │ ├── content.rst
│ │ │ └── rn.rst
│ │ ├── tensorboard-neuron.rst
│ │ ├── tensorflow/
│ │ │ ├── tensorflow-modelserver-neuron/
│ │ │ │ ├── tensorflow-modelserver-neuron-v2.rst
│ │ │ │ ├── tensorflow-modelserver-neuron.rst
│ │ │ │ └── tensorflow-modelserver-neuronx.rst
│ │ │ ├── tensorflow-neuron/
│ │ │ │ ├── tensorflow-neuron-v2.rst
│ │ │ │ └── tensorflow-neuron.rst
│ │ │ └── tensorflow-neuronx/
│ │ │ └── tensorflow-neuronx.rst
│ │ └── torch-neuron.rst
│ ├── components/
│ │ ├── compiler.rst
│ │ ├── containers.rst
│ │ ├── dev-tools.rst
│ │ ├── dlamis.rst
│ │ ├── index.rst
│ │ ├── jax.rst
│ │ ├── nki-lib.rst
│ │ ├── nki.rst
│ │ ├── nxd-core.rst
│ │ ├── nxd-inference.rst
│ │ ├── nxd-training.rst
│ │ ├── pytorch.rst
│ │ └── runtime.rst
│ ├── documentation/
│ │ └── neuron-documentation.rst
│ ├── index.rst
│ ├── prev/
│ │ ├── 2.25.0/
│ │ │ ├── compiler.rst
│ │ │ ├── containers.rst
│ │ │ ├── dlami.rst
│ │ │ ├── docs-and-samples.rst
│ │ │ ├── index.rst
│ │ │ ├── nx-jax.rst
│ │ │ ├── nx-pytorch.rst
│ │ │ ├── nxd-core.rst
│ │ │ ├── nxd-inference.rst
│ │ │ ├── nxd-training.rst
│ │ │ ├── runtime.rst
│ │ │ └── tools.rst
│ │ ├── 2.26.0/
│ │ │ ├── containers.rst
│ │ │ ├── dlami.rst
│ │ │ ├── index.rst
│ │ │ ├── nki.rst
│ │ │ ├── nx-jax.rst
│ │ │ ├── nx-pytorch.rst
│ │ │ ├── nxd-core.rst
│ │ │ ├── nxd-inference.rst
│ │ │ ├── runtime.rst
│ │ │ └── tools.rst
│ │ ├── 2.26.1.rst
│ │ ├── 2.27.0/
│ │ │ ├── compiler.rst
│ │ │ ├── containers.rst
│ │ │ ├── dlami.rst
│ │ │ ├── index.rst
│ │ │ ├── nki-lib.rst
│ │ │ ├── nki.rst
│ │ │ ├── nx-pytorch.rst
│ │ │ ├── nxd-inference.rst
│ │ │ ├── runtime.rst
│ │ │ └── tools.rst
│ │ ├── 2.27.1.rst
│ │ ├── 2.28.0.rst
│ │ ├── 2.28.1.rst
│ │ ├── content.rst
│ │ └── rn.rst
│ └── releasecontent.rst
├── requirements-python310.txt
├── requirements-python38.txt
├── requirements.txt
├── setup/
│ ├── index.rst
│ ├── index.txt-back
│ ├── install-templates/
│ │ ├── al2-python.rst
│ │ ├── inf1/
│ │ │ ├── compile_mode.rst
│ │ │ ├── deploy_mode.rst
│ │ │ ├── develop_mode.rst
│ │ │ ├── dlami-enable-neuron-mxnet.rst
│ │ │ ├── dlami-enable-neuron-pytorch.rst
│ │ │ ├── launch-inf1-ami.rst
│ │ │ ├── launch-inf1-dlami-aws-cli.rst
│ │ │ ├── launch-inf1-dlami.rst
│ │ │ ├── neuron-pip-install.rst
│ │ │ ├── neuron-pip-setup.rst
│ │ │ ├── note-setup-cntr.rst
│ │ │ ├── note-setup-general.rst
│ │ │ ├── note-setup-libnrt-warning.rst
│ │ │ └── tensorboard-plugin-neuron-pip-install.rst
│ │ ├── inf2/
│ │ │ ├── dlami-enable-neuron-pytorch.rst
│ │ │ ├── launch-inf2-dlami.rst
│ │ │ └── note-setup-libnrt-warning.rst
│ │ ├── launch-instance.txt
│ │ ├── launch-trn1-dlami.rst
│ │ ├── trn1/
│ │ │ └── dlami-notes.rst
│ │ └── trn1-ga-warning.txt
│ ├── jax/
│ │ ├── dlami.rst
│ │ ├── dlc.rst
│ │ ├── index.rst
│ │ └── manual.rst
│ ├── jax-neuronx.rst
│ ├── legacy-inf1/
│ │ ├── index.rst
│ │ └── pytorch.rst
│ ├── multiframework-dlami.rst
│ ├── mxnet-neuron.rst
│ ├── notebook/
│ │ ├── running-jupyter-notebook-as-script.rst
│ │ └── setup-jupyter-notebook-steps-troubleshooting.rst
│ ├── pytorch/
│ │ ├── dlami.rst
│ │ ├── dlc.rst
│ │ ├── index.rst
│ │ ├── manual.rst
│ │ ├── update-dlami.rst
│ │ ├── update-dlc.rst
│ │ └── update-manual.rst
│ ├── setup-rocky-linux-9.rst
│ ├── setup-troubleshooting.rst
│ ├── torch-neuron-ubuntu20.rst
│ ├── torch-neuron.rst
│ ├── torch-neuronx.rst
│ └── troubleshooting.rst
├── src/
│ ├── benchmark/
│ │ ├── helper_scripts/
│ │ │ ├── llmperf_dp.patch
│ │ │ ├── llmperf_reasoning.patch
│ │ │ └── neuron_perf.patch
│ │ └── tensorflow/
│ │ ├── distilbert-base-uncased-finetuned-sst-2-english_benchmark.py
│ │ └── distilbert-base-uncased-finetuned-sst-2-english_compile.py
│ ├── examples/
│ │ ├── mxnet/
│ │ │ ├── README.md
│ │ │ ├── data_parallel/
│ │ │ │ ├── benchmark_utils.py
│ │ │ │ ├── data_parallel_tutorial.ipynb
│ │ │ │ └── parallel.py
│ │ │ ├── mxnet-gluon-tutorial.ipynb
│ │ │ ├── resnet50/
│ │ │ │ └── resnet50.ipynb
│ │ │ └── resnet50_neuroncore_groups.ipynb
│ │ ├── neuron-monitor/
│ │ │ └── neuron-monitor-grafana.json
│ │ ├── pytorch/
│ │ │ ├── bert_tutorial/
│ │ │ │ ├── README.md
│ │ │ │ ├── THIRD
│ │ │ │ ├── THIRD PARTY LICENSE.txt
│ │ │ │ ├── bert_benchmark_utils.py
│ │ │ │ ├── glue_mrpc_dev.tsv
│ │ │ │ ├── parallel.py
│ │ │ │ ├── tutorial_pretrained_bert.ipynb
│ │ │ │ └── tutorial_pretrained_bert_shared_weights.ipynb
│ │ │ ├── byoc_sm_bert_tutorial/
│ │ │ │ ├── code/
│ │ │ │ │ └── inference.py
│ │ │ │ ├── container/
│ │ │ │ │ └── Dockerfile
│ │ │ │ └── sagemaker_container_neuron.ipynb
│ │ │ ├── libtorch_demo/
│ │ │ │ ├── bert_neuronx/
│ │ │ │ │ ├── compile.py
│ │ │ │ │ └── detect_instance.py
│ │ │ │ ├── clean.sh
│ │ │ │ ├── example_app/
│ │ │ │ │ ├── README.txt
│ │ │ │ │ ├── build.sh
│ │ │ │ │ ├── core_count.hpp
│ │ │ │ │ ├── example_app.cpp
│ │ │ │ │ ├── utils.cpp
│ │ │ │ │ └── utils.hpp
│ │ │ │ ├── neuron.patch
│ │ │ │ ├── run_tests.sh
│ │ │ │ ├── setup.sh
│ │ │ │ ├── tokenizers_binding/
│ │ │ │ │ ├── build.sh
│ │ │ │ │ ├── remote_rust_tokenizer.h
│ │ │ │ │ ├── run.sh
│ │ │ │ │ ├── run_python.sh
│ │ │ │ │ ├── tokenizer_test
│ │ │ │ │ ├── tokenizer_test.cpp
│ │ │ │ │ └── tokenizer_test.py
│ │ │ │ └── trace_bert_neuron.py
│ │ │ ├── mnist_mlp/
│ │ │ │ ├── train_monitor.py
│ │ │ │ └── train_tb.py
│ │ │ ├── neuronx_distributed/
│ │ │ │ └── t5-inference/
│ │ │ │ ├── t5-inference-tutorial.ipynb
│ │ │ │ ├── t5_model_layers.py
│ │ │ │ ├── t5_models.py
│ │ │ │ └── wrapper.py
│ │ │ ├── pipeline_tutorial/
│ │ │ │ └── neuroncore_pipeline_pytorch.ipynb
│ │ │ ├── resnet50.ipynb
│ │ │ ├── resnet50_partition.ipynb
│ │ │ ├── torch-neuronx/
│ │ │ │ ├── bert-base-cased-finetuned-mrpc-inference-on-trn1-tutorial.ipynb
│ │ │ │ ├── resnet50-inference-on-trn1-tutorial.ipynb
│ │ │ │ └── t5-inference-tutorial.ipynb
│ │ │ ├── torchserve/
│ │ │ │ ├── benchmark_bert.py
│ │ │ │ ├── config.json
│ │ │ │ ├── handler_bert.py
│ │ │ │ ├── handler_bert_neuronx.py
│ │ │ │ ├── infer_bert.py
│ │ │ │ ├── torchserve.config
│ │ │ │ ├── trace_bert_neuron.py
│ │ │ │ └── trace_bert_neuronx.py
│ │ │ ├── transformers-marianmt.ipynb
│ │ │ └── yolo_v4.ipynb
│ │ └── tensorflow/
│ │ ├── bert_demo/
│ │ │ ├── LICENSE
│ │ │ ├── README.md
│ │ │ ├── bert_client.py
│ │ │ ├── bert_model.py
│ │ │ ├── bert_model_server.py
│ │ │ ├── bert_no_model.py
│ │ │ ├── bert_server.py
│ │ │ ├── download_mrpc_data.py
│ │ │ ├── glue_mrpc_dev.tsv
│ │ │ ├── latency_printer.py
│ │ │ ├── mrpc.proto
│ │ │ ├── mrpc_feature.py
│ │ │ ├── mrpc_pb2.py
│ │ │ ├── mrpc_pb2_grpc.py
│ │ │ ├── protoc.sh
│ │ │ ├── setup.py
│ │ │ ├── tokenization.py
│ │ │ ├── tune_save.sh
│ │ │ └── uncased_L-24_H-1024_A-16.vocab.txt
│ │ ├── huggingface_bert/
│ │ │ └── huggingface_bert.ipynb
│ │ ├── k8s_bert_demo/
│ │ │ ├── Dockerfile.tfserving_example
│ │ │ ├── README.md
│ │ │ ├── bert_client.py
│ │ │ └── bert_service.yml
│ │ ├── keras_resnet50/
│ │ │ ├── LICENSE
│ │ │ ├── README.md
│ │ │ ├── fp32tofp16.py
│ │ │ ├── full_sweep
│ │ │ ├── gen_resnet50_keras.py
│ │ │ ├── infer_resnet50_keras.py
│ │ │ ├── infer_resnet50_keras_loadtest.py
│ │ │ ├── keras_resnet50.ipynb
│ │ │ ├── optimize_for_inference.py
│ │ │ ├── pb2sm_compile.py
│ │ │ └── run_all
│ │ ├── openpose_demo/
│ │ │ └── openpose.ipynb
│ │ ├── ssd300_demo/
│ │ │ ├── README.md
│ │ │ ├── ssd300_detection.py
│ │ │ ├── ssd300_evaluation.py
│ │ │ ├── ssd300_evaluation_client.py
│ │ │ └── ssd300_model.py
│ │ ├── tensorflow-neuronx/
│ │ │ └── tfneuronx-roberta-base-tutorial.ipynb
│ │ ├── tensorflow_resnet50/
│ │ │ └── resnet50.ipynb
│ │ ├── tensorflow_serving_tutorial.rst
│ │ ├── yolo_v3_demo/
│ │ │ ├── yolo_v3.ipynb
│ │ │ └── yolo_v3_coco_saved_model.py
│ │ └── yolo_v4_demo/
│ │ ├── README.md
│ │ ├── evaluate.ipynb
│ │ └── yolo_v4_coco_saved_model.py
│ ├── helperscripts/
│ │ ├── installationScripts/
│ │ │ └── python_instructions.txt
│ │ ├── n2-helper.py
│ │ ├── n2-manifest.json
│ │ ├── neuron-releases-manifest.json
│ │ ├── neuron-setup-example.py
│ │ ├── neuronsetuphelper.py
│ │ └── release-manifest-def.py
│ ├── k8/
│ │ ├── bert_service.yml
│ │ ├── k8s-neuron-device-plugin-rbac.yml
│ │ ├── k8s-neuron-device-plugin.yml
│ │ ├── k8s-neuron-monitor-daemonset.yml
│ │ ├── k8s-neuron-scheduler-configmap.yml
│ │ ├── k8s-neuron-scheduler-eks.yml
│ │ ├── k8s-neuron-scheduler.yml
│ │ ├── k8s-ultraserver-init-script.sh
│ │ ├── my-scheduler.yml
│ │ └── neuron-problem-detector/
│ │ ├── k8s-neuron-problem-detector-and-recovery-config.yml
│ │ ├── k8s-neuron-problem-detector-and-recovery-rbac.yml
│ │ └── k8s-neuron-problem-detector-and-recovery.yml
│ ├── libnrt/
│ │ ├── README.md
│ │ └── include/
│ │ ├── ndl/
│ │ │ ├── ndl.h
│ │ │ ├── neuron_driver_shared.h
│ │ │ └── neuron_driver_shared_tensor_batch_op.h
│ │ └── nrt/
│ │ ├── ndebug_stream.h
│ │ ├── nds/
│ │ │ └── neuron_ds.h
│ │ ├── nec.h
│ │ ├── nrt.h
│ │ ├── nrt_async.h
│ │ ├── nrt_async_sendrecv.h
│ │ ├── nrt_experimental.h
│ │ ├── nrt_profile.h
│ │ ├── nrt_status.h
│ │ ├── nrt_sys_trace.h
│ │ └── nrt_version.h
│ ├── neuron-gatherinfo/
│ │ ├── LICENSE
│ │ ├── clear_params_tfpb.py
│ │ ├── mx_neuron_check_model.py
│ │ ├── neuron-gatherinfo.py
│ │ └── tf_neuron_check_model.py
│ └── neuronperf/
│ ├── LICENSE
│ ├── README.md
│ ├── build.sh
│ ├── conf.py
│ ├── model_neuron_b1.csv
│ ├── pyproject.toml
│ ├── src/
│ │ └── neuronperf/
│ │ ├── __init__.py
│ │ ├── __version__.py
│ │ ├── benchmarking.py
│ │ ├── compile_constants.py
│ │ ├── cpu/
│ │ │ ├── __init__.py
│ │ │ └── cpu.py
│ │ ├── logging.py
│ │ ├── model_index.py
│ │ ├── mxnet/
│ │ │ ├── __init__.py
│ │ │ └── mxnet.py
│ │ ├── py.typed
│ │ ├── reporting.py
│ │ ├── scripts/
│ │ │ ├── __init__.py
│ │ │ └── run_benchmark_file.py
│ │ ├── tensorflow/
│ │ │ ├── __init__.py
│ │ │ └── tensorflow.py
│ │ ├── timing.py
│ │ └── torch/
│ │ ├── __init__.py
│ │ └── torch.py
│ └── test/
│ └── test_neuronperf.py
├── static/
│ ├── google673a8c4fbaa024d8.html
│ ├── robots.txt
│ └── sitemap1.xml
└── tools/
├── index.rst
├── neuron-explorer/
│ ├── get-started.rst
│ ├── how-to-link-view-source-code.rst
│ ├── how-to-profile-workload.rst
│ ├── index.rst
│ ├── migration-faq.rst
│ ├── overview-ai-recommendations.rst
│ ├── overview-database-viewer.rst
│ ├── overview-device-profiles.rst
│ ├── overview-hierarchy-view.rst
│ ├── overview-memory-viewer.rst
│ ├── overview-summary-page.rst
│ ├── overview-system-profiles.rst
│ ├── overview-tensor-viewer.rst
│ └── view-perfetto.rst
├── neuron-sys-tools/
│ ├── index.rst
│ ├── nccom-test.rst
│ ├── neuron-ls.rst
│ ├── neuron-monitor-user-guide.rst
│ ├── neuron-sysfs-user-guide.rst
│ └── neuron-top-user-guide.rst
├── profiler/
│ ├── neuron-profile-user-guide.rst
│ └── neuron-profiler-2-0-beta-user-guide.rst
├── tensorboard/
│ ├── getting-started-tensorboard-neuronx-plugin.rst
│ └── index.rst
├── third-party-solutions.rst
└── tutorials/
├── index.rst
├── performance-profiling-vllm.rst
├── torch-neuronx-profiling-with-tb.rst
├── tutorial-neuron-monitor-mnist.rst
└── tutorial-tensorboard-scalars-mnist.rst
================================================
FILE CONTENTS
================================================
================================================
FILE: .github/ISSUE_TEMPLATE/bug-report.yml
================================================
---
name: "🐛 Bug Report"
description: Report a bug
title: "(short issue description)"
labels: [bug, needs-triage]
assignees: []
body:
- type: textarea
id: description
attributes:
label: Describe the bug
description: What is the problem? Provide a clear description of your issue and the steps you took that produced it.
validations:
required: true
- type: textarea
id: modelname
attributes:
label: Model Name
description: Provide Model Name
validations:
required: true
- type: textarea
id: workloadtype
attributes:
label: Describe the workload type
description: Note the type of workload (such as Inference or Training) and any specific details about the workload configuration.
validations:
required: true
- type: textarea
id: instancetype
attributes:
label: Instance Type
description: |
Provide the AWS EC2 instance type you used to run the workload (such as `inf2.xlarge`, `trn1.32xlarge`, `trn2.48xlarge` etc.)
validations:
required: true
- type: textarea
id: release
attributes:
label: Release version
description: |
Provide the Neuron SDK release version (such as `2.25.0`) you are using, and all relevant Neuron component versions.
```
apt list --installed | grep -i -e neuron
pip list | grep -i -e neuron -e torch -e transformers -e jax
```
- type: textarea
id: reproduction
attributes:
label: Reproduction Steps
description: |
Provide the type of the model and links to any tutorials you may have used, as additional context.
Provide a self-contained, concise snippet of code that can be used to reproduce the issue.
For more complex issues provide a repo with the smallest sample that reproduces the bug.
Avoid including business logic or unrelated code as it makes diagnosis more difficult.
The code sample should be an SSCCE. See http://sscce.org/ for details. In short, please provide a code sample that we can copy/paste, run and reproduce.
validations:
required: true
- type: checkboxes
id: regression
attributes:
label: Regression Issue
description: Is this as regression (did it work in a previous version and not now)?
If this is a regression, provide the Neuron SDK release version where this configuration worked for you.
options:
- label: Select this option if this issue appears to be a regression.
required: false
- type: textarea
id: solution
attributes:
label: Possible Solution
description: |
Suggest a fix or reason for the bug, if you know:
validations:
required: false
- type: textarea
id: context
attributes:
label: Logs/Context/Additional Information
description: |
Anything else that might be relevant for troubleshooting this bug. Providing context helps us come up with a solution that is most useful in the real world. When applicable, please provide HLOs and compiler commands.
validations:
required: false
================================================
FILE: .github/ISSUE_TEMPLATE/config.yml
================================================
blank_issues_enabled: false
================================================
FILE: .github/ISSUE_TEMPLATE/documentation.yml
================================================
---
name: "📕 Documentation Issue"
description: Report an issue in the documentation and Developer Guide
title: "(short issue description)"
labels: [documentation, needs-triage]
assignees: []
body:
- type: textarea
id: description
attributes:
label: Describe the issue
description: A clear and concise description of the issue.
validations:
required: true
- type: textarea
id: links
attributes:
label: Links
description: |
Include links to affected documentation page(s).
validations:
required: true
================================================
FILE: .github/ISSUE_TEMPLATE/feature-request.yml
================================================
---
name: 🚀 Feature Request
description: Suggest an idea for this project
title: "(short issue description)"
labels: [feature-request, needs-triage]
assignees: []
body:
- type: textarea
id: description
attributes:
label: Describe the feature
description: A clear and concise description of the feature you are proposing.
validations:
required: true
- type: textarea
id: use-case
attributes:
label: Use Case
description: |
Why do you need this feature?
validations:
required: true
- type: textarea
id: solution
attributes:
label: Proposed Solution
description: |
Provide detailed suggestions or requirements for this proposed feature. If you have them, include any reference implementation details (or even links to prototypes).
validations:
required: false
- type: textarea
id: other
attributes:
label: Other Information
description: |
Any additional details or information you can provide, including links to related content or similar issues.
validations:
required: false
- type: checkboxes
id: ack
attributes:
label: Acknowledgements
options:
- label: I may be able to implement this feature request
required: false
================================================
FILE: .github/pull_request_template.md
================================================
**IMPORTANT!** _If this is a documentation PR for a specific release, this PR must go the corresponding release branch_ (`release-X.XX.X`). _If it is an "out-of-band" doc update, the PR must go to the_ `master` _branch_.
## Required PR information
To expedite approvals and merges for releases, provide the following information (select the `...` button to the right at the top of your PR message to edit it):
> **AWS email alias**: {_your-name_}@amazon.com
>**Description**: {_What this documentation change is and why you made it. If you have a corresponding Jira ticket or content plan, link it here. The more details you provide around any decisions you made when preparing the docs, the less annoying comments you'll get preparing to release it._}
> **Date this must be published by**: {_If empty, we will assume the release date for the branch you're merging into._}
> **Link to ReadTheDocs staging for this branch's doc changes**: https://awsdocs-neuron-staging.readthedocs-hosted.com/en/{YOUR_BRANCH_NAME_HERE}/
> **Set the `docs-review-needed` label on the PR for tracking.**
## Before you request approvals
> Run a spelling and grammar check over your prose and make the changes it suggests. VSCode has a number of extensions (cSpell, LTeX) that you can use. You can also provide the rendered HTML for (or a cut-and-paste of) your pages to an AI and have it correct your spelling, grammar, and formatting issues. If you need an advanced prompt, contact @erickson-doug.
## Approvers
We require 3-4 approvers to merge for non-trivial content changes (where a "trivial" change is a typo/grammar fix or a minor update to the format syntax):
1. A senior+ engineer who will review your documentation for technical accuracy and clarity in communicating the technical concepts in your work
2. A product manager for your Neuron component area who will review it for customer relevance and product/component/feature messaging
3. The lead tech writer (@erickson-doug) who will review your work for overall doc design and quality, and perform the merge when all approvals are met
4. (For PRs with code/notebook submissions) A QA/test engineer who can run your code and confirm the results.
Make sure you get a commitment from these reviewers in advance! It's hard to get good quality doc reviews in order in the 11th hour of a release.
**Note**: For trivial changes, you only need @erickson-doug's approval. He will merge your content once he's confirmed the fixes on staging.
## Doc review checklist
### Engineering reviewer checklist
- [ ] I've confirmed that the contributions in this PR meet the current [AWS Neuron writing guidelines](https://quip-amazon.com/m97CAO0kQFEU/Writing-for-AWS-Neuron).
- [ ] I've confirmed that the documentation submitted is technically correct to the best of my knowledge.
- [ ] I've confirmed that the documentation submitted has no spelling or grammar errors or use of internal jargon/terminology.
- [ ] I've verified the changes render correctly on RTD (link above).
- [ ] (If code is included) I've run tests to verify the contents of the change.
---
## For PRs that include code or notebook examples
**MANDATORY: PR must include test run output**
Provide this information for the QA reviewer in order to expedite their review.
**Test run output:**
Specify the release version, instance size and type, OS type and test output.
**For Training tutorials:**
{Convergence graph for training tutorials}
{Performance metrics `average_throughput`, `latency_p50`, `latency_p99` and MFU% if available}
Make sure this PR contains correct classification terms (Alpha, Beta, and Stable).
If possible, provide your results or a link to them for the reviewer to check your work.
## Code example/notebook content PR checklist
- [ ] (If applicable) I've automated a test to safeguard my changes from regression.
- [ ] (If applicable) I've posted test collateral to prove my change was effective and not harmful.
- [ ] (If applicable) I've added someone from QA to the list of reviewers. Do this if you didn't make an automated test or feel it's appropriate for another reason.
- [ ] (If applicable) I've reviewed the licenses of updated and new binaries and their dependencies to make sure all licenses are on the pre-approved Amazon license list. See https://inside.amazon.com/en/services/legal/us/OpenSource/Pages/BlessedOpenSourceLicenses.aspx.
By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.
================================================
FILE: .github/stale_issue_mark_close_workflow.yml
================================================
name: Close inactive issues
on:
schedule:
- cron: "30 1 * * *"
jobs:
close-issues:
runs-on: ubuntu-latest
permissions:
issues: write
pull-requests: write
steps:
- uses: actions/stale@v5
with:
days-before-issue-stale: 30
days-before-issue-close: 14
stale-issue-label: "stale"
stale-issue-message: "This issue is stale because it has been open for 30 days with no activity."
close-issue-message: "This issue was closed because it has been inactive for 14 days since being marked as stale."
days-before-pr-stale: -1
days-before-pr-close: -1
repo-token: ${{ secrets.GITHUB_TOKEN }}
================================================
FILE: .github/workflows/acknowledge-new-issue.yml
================================================
name: Acknowledge New Issue
on:
issues:
types: [opened]
permissions:
issues: write
jobs:
acknowledge:
runs-on: ubuntu-latest
steps:
- name: Comment on issue
uses: actions/github-script@v7
with:
script: |
const creator = context.payload.issue.user.login;
await github.rest.issues.createComment({
owner: context.repo.owner,
repo: context.repo.repo,
issue_number: context.payload.issue.number,
body: `Hi @${creator}, Thank you for filing the issue! We will take a look and get back to you.`
});
================================================
FILE: .github/workflows/auto-label-issues.yml
================================================
# Auto-label issues based on content keywords
name: auto-label-issues
on:
issues:
types: [opened]
jobs:
auto-label-issues:
runs-on: ubuntu-latest
permissions:
issues: write
steps:
- name: Analyze issue content
id: analyze_content
uses: actions/github-script@v7
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
ISSUE_TITLE: ${{ github.event.issue.title }}
ISSUE_BODY: ${{ github.event.issue.body }}
with:
script: |
const title = process.env.ISSUE_TITLE || '';
const body = process.env.ISSUE_BODY || '';
const content = `${title} ${body}`;
const labels = [];
// =============================================================================
// LABEL CONFIGURATION - Easy to update dictionary
// Add keywords, typos, or synonyms to the arrays below
// =============================================================================
const labelConfig = {
// ----- Issue Type Labels (mutually exclusive) -----
bug: {
keywords: [
// Standard terms
'bug', 'error', 'crash', 'fail', 'failed', 'failure', 'failing',
'broken', 'exception', 'traceback', 'segfault', 'segmentation fault',
// Synonyms
'issue', 'problem', 'defect', 'fault', 'glitch', 'malfunction',
'wrong', 'incorrect', 'unexpected',
'hang', 'hanging', 'hung', 'freeze', 'frozen',
'timeout', 'timed out',
'oom', 'out of memory', 'memory error',
'nan', 'diverge', 'diverged',
// Common typos
'bugg', 'bgu', 'eror', 'errror', 'crahs', 'fial', 'brokn', 'broke'
],
patterns: [/not\s*work/i, /doesn'?t\s*work/i, /won'?t\s*work/i, /can'?t\s*work/i]
},
documentation: {
keywords: [
// Standard terms
'doc', 'docs', 'documentation', 'readme',
'guide', 'tutorial', 'howto', 'how-to', 'how to',
'typo', 'typos', 'spelling', 'grammar',
'example', 'examples', 'sample', 'samples',
'instruction', 'instructions',
'clarify', 'clarification', 'unclear', 'confusing',
'outdated', 'out of date', 'stale',
'missing documentation', 'missing docs',
'broken link', 'dead link', '404',
// Common typos
'documention', 'documenation', 'documentaion', 'tutoral', 'toturial'
],
patterns: [/issue\s*on\s*page/i, /page\s*.*\.html/i]
},
'feature-request': {
keywords: [
// Standard terms
'feature', 'feature request', 'feature-request',
'enhancement', 'improvement',
'implement', 'implementation',
'new feature', 'add feature',
'support for', 'add support',
'would be nice', 'would be great', 'would be helpful',
'suggestion', 'suggest', 'proposal', 'propose',
'wishlist', 'wish list',
// Common typos
'feture', 'featrue', 'enchancement', 'improvment'
],
patterns: [/add\s+support\s+for/i, /please\s+add/i, /would\s+be\s+(nice|great|helpful)/i]
},
// ----- Hardware Labels (independent - multiple can be applied) -----
Trn1: {
keywords: [
'trn1', 'trn-1', 'trn 1', 'trn1n',
'trn1.2xlarge', 'trn1.32xlarge', 'trn1n.32xlarge',
'trainium', 'trainium1', 'trainium 1', 'trainium-1',
// Common typos
'tranium', 'trainuim', 'trn-1n'
],
patterns: [/trn1n?(?:\.[0-9]*xlarge)?/i, /trainium\s*1?(?!\s*2)/i]
},
Trn2: {
keywords: [
'trn2', 'trn-2', 'trn 2',
'trn2.48xlarge',
'trainium2', 'trainium 2', 'trainium-2',
// Common typos
'tranium2', 'trainuim2'
],
patterns: [/trn2(?:\.[0-9]*xlarge)?/i, /trainium\s*2/i]
},
Inf1: {
keywords: [
'inf1', 'inf-1', 'inf 1',
'inf1.xlarge', 'inf1.2xlarge', 'inf1.6xlarge', 'inf1.24xlarge',
'inferentia', 'inferentia1', 'inferentia 1', 'inferentia-1',
// Common typos
'infertia', 'inferntia', 'infernita'
],
patterns: [/inf1(?:\.[0-9]*xlarge)?/i, /inferentia\s*1?(?!\s*2)/i]
},
Inf2: {
keywords: [
'inf2', 'inf-2', 'inf 2',
'inf2.xlarge', 'inf2.8xlarge', 'inf2.24xlarge', 'inf2.48xlarge',
'inferentia2', 'inferentia 2', 'inferentia-2',
// Common typos
'infertia2', 'inferntia2', 'infernita2'
],
patterns: [/inf2(?:\.[0-9]*xlarge)?/i, /inferentia\s*2/i]
},
// ----- Use Case Labels (independent - both can be applied) -----
Inference: {
keywords: [
// Standard terms
'inference', 'inferencing',
'predict', 'prediction', 'predictions', 'predicting',
'serving', 'serve', 'server',
'batch inference', 'real-time', 'realtime',
'endpoint', 'endpoints',
// Common typos
'infernce', 'inferance', 'prediciton', 'deploymnet'
],
patterns: [/infer(?:ence|ring)?/i, /predict(?:ion|ing)?/i, /deploy(?:ment|ing)?/i]
},
Training: {
keywords: [
// Standard terms
'training', 'train', 'trained',
'fine-tune', 'finetune', 'fine tune', 'finetuning', 'fine-tuning',
'pretrain', 'pre-train', 'pretraining', 'pre-training',
'learning', 'learn',
'gradient', 'gradients',
'backward', 'backprop', 'backpropagation',
'loss', 'convergence', 'converge',
'epoch', 'epochs',
'checkpoint', 'checkpointing',
// Common typos
'trainig', 'traning', 'trainin', 'fintune', 'finetunning'
],
patterns: [/train(?:ing|ed)?/i, /fine[\s-]?tun(?:e|ing)/i, /pre[\s-]?train(?:ing)?/i]
}
};
// =============================================================================
// MATCHING LOGIC
// =============================================================================
function matchesLabel(config) {
const contentLower = content.toLowerCase();
// Check keywords (case-insensitive substring match)
for (const keyword of config.keywords) {
if (contentLower.includes(keyword.toLowerCase())) {
return true;
}
}
// Check regex patterns
for (const pattern of config.patterns) {
if (pattern.test(content)) {
return true;
}
}
return false;
}
// Issue Type Labels - MUTUALLY EXCLUSIVE (priority: bug > documentation > feature-request)
if (matchesLabel(labelConfig.bug)) {
labels.push('bug');
} else if (matchesLabel(labelConfig.documentation)) {
labels.push('documentation');
} else if (matchesLabel(labelConfig['feature-request'])) {
labels.push('feature-request');
}
// Hardware/Instance Type Labels - INDEPENDENT (multiple can be applied)
if (matchesLabel(labelConfig.Trn1)) {
labels.push('Trn1');
}
if (matchesLabel(labelConfig.Trn2)) {
labels.push('Trn2');
}
if (matchesLabel(labelConfig.Inf1)) {
labels.push('Inf1');
}
if (matchesLabel(labelConfig.Inf2)) {
labels.push('Inf2');
}
// Use Case Labels - INDEPENDENT (both can be applied)
if (matchesLabel(labelConfig.Inference)) {
labels.push('Inference');
}
if (matchesLabel(labelConfig.Training)) {
labels.push('Training');
}
core.setOutput('labels', labels.join(','));
core.setOutput('has_labels', labels.length > 0);
- name: Apply labels to issue
if: steps.analyze_content.outputs.has_labels == 'true'
env:
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
run: |
IFS=',' read -ra LABELS <<< "${{ steps.analyze_content.outputs.labels }}"
for label in "${LABELS[@]}"; do
gh issue edit ${{ github.event.issue.number }} --add-label "$label" -R ${{ github.repository }}
done
================================================
FILE: .gitignore
================================================
_build/
__pycache__/
.venv/
.DS_Store
src/examples/pytorch/libtorch_demo.tar.gz
src/neuronperf.tar.gz
*-checkpoint.ipynb
.idea/
.vscode/
nki/*/generated/
uncommitted/
================================================
FILE: .readthedocs.yml
================================================
# .readthedocs.yml
# Read the Docs configuration file
# See https://docs.readthedocs.io/en/stable/config-file/v2.html for details
# Required
version: 2
# Set the version of Python and other tools you might need
build:
os: "ubuntu-22.04"
tools:
python: "3.10"
# jobs:
# pre_build:
# - python -m sphinx -b linkcheck . _build/linkcheck
# Build documentation in the docs/ directory with Sphinx
sphinx:
configuration: conf.py
#conda
#conda:
# file: readthedocs-environment.yml
# Build documentation with MkDocs
#mkdocs:
# configuration: mkdocs.yml
# Optionally build your docs in additional formats such as PDF
#formats:
# - pdf
# Optionally set the version of Python and requirements required to build your docs
python:
install:
- requirements: requirements.txt
================================================
FILE: CODEOWNERS
================================================
# This file creates codeowners for the documentation. It will allow setting code reviewers for all Pull requests to merge to the master branch
# Each line is a file pattern followed by one or more owners.
# Reference guide - https://docs.github.com/en/github/creating-cloning-and-archiving-repositories/creating-a-repository-on-github/about-code-owners#example-[…]ners-file
# Example - These owners will be the default owners for everything in
# the repo. Unless a later match takes precedence,
# @global-owner1 and @global-owner2 will be requested for
# review when someone opens a pull request.
# * @global-owner1 @global-owner2
* @aws-maens @micwade-aws @musunita @aws-sadaf @rgrandhiamzn @eshalakhotia @jluntamazon @jeffhataws @aws-rhsoln @hannanjgaws @PrashantSaraf @aws-donkrets @aws-singhada @gsnaws @awsjoshir @sidjoshiaws @pinak-p @vikas-paliwal-aws @aarondou @mrinalks @erickson-doug @lnixaws @micwade-aws
src/examples/mxnet/ @aws-rhsoln @aws-sadaf @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
neuron-guide/neuron-frameworks/mxnet-neuron/ @aws-rhsoln @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
neuron-guide/neuron-frameworks/mxnet-neuron/tutorials/ @musunita @aws-rhsoln @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
src/examples/tensorflow/ @awshaichen @aws-sadaf @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
neuron-guide/neuron-frameworks/tensorflow-neuron/ @awshaichen @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
neuron-guide/neuron-frameworks/tensorflow-neuron/tutorials/ @musunita @awshaichen @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
src/examples/pytorch/ @jluntamazon @aws-sadaf @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
neuron-guide/neuron-frameworks/pytorch-neuron/ @jluntamazon @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
neuron-guide/neuron-frameworks/pytorch-neuron/tutorials/ @musunita @jluntamazon @aws-maens @vikas-paliwal-aws @rgrandhiamzn @eshalakhotia
libraries/nxd-inference/ @huntingcarlisle @lccasagrande @lipovsek-aws @erickson-doug @eshalakhotia @pinak-p @hannanjgaws @akhil-aws @ahimsh-aws @rgrandhiamzn @yahavb @FThompsonAWS @gsnaws @sidjoshiaws @jluntamazon @musunita
================================================
FILE: CONTRIBUTING.md
================================================
# Contributing Guidelines
Thank you for your interest in contributing to our project. Whether it's a bug report, new feature, correction, or additional
documentation, we greatly value feedback and contributions from our community.
Please read through this document before submitting any issues or pull requests to ensure we have all the necessary
information to effectively respond to your bug report or contribution.
## Reporting Bugs/Feature Requests
We welcome you to use the GitHub issue tracker to report bugs or suggest features.
When filing an issue, please check existing open, or recently closed, issues to make sure somebody else hasn't already
reported the issue. Please try to include as much information as you can. Details like these are incredibly useful:
* A reproducible test case or series of steps
* The version of our code being used
* Any modifications you've made relevant to the bug
* Anything unusual about your environment or deployment
## Contributing Workflow (via Pull Requests)
Contributions via pull requests are much appreciated. Before sending us a pull request, please ensure that:
1. You are working against the latest source on the *master* branch.
2. You check existing open, and recently merged, pull requests to make sure someone else hasn't addressed the problem already.
3. You open an issue to discuss any significant work - we would hate for your time to be wasted.
**Important**: Currently, local doc builds require a Python 3.9 environment. If you are on MacOS, you can install it from the terminal with `brew install python@3.9`. Add it to your working path with `brew link python@3.9` and confirm it works by running `python3.9 --version`.
### Docker Build
If you don't have Python 3.9/3.10 or a compatible gcc toolchain, use the Docker workflow:
```bash
./build.sh build # Build Docker image (first time only)
./build.sh html # Build HTML docs to _build/html/
./build.sh shell # Interactive shell for debugging
./build.sh clean # Remove _build/ directory
```
### Manual Build
To send us a pull request, please:
1. Clone the repository locally:
```bash
git clone git@github.com:YOUR-USERNAME/private-aws-neuron-sdk-staging.git
```
2. Install the build dependencies. This requires a Python 3.9 installation and venv:
```bash
cd .. # The root folder where you have your cloned Git repos; don't run this in the repo folder but one level up or you'll have venv files in your repo folder
python3.9 -m venv venv && . venv/bin/activate
pip install -U pip
cd private-aws-neuron-sdk-staging
pip install -r requirements.txt
```
3. Build the documentation into HTML. This command will allow you to view the
rendered documentation by opening the generated `_build/html/index.html`. On first run, this will take about 15 mins. Subsequent html generations are incremental and will take less time.
Run:
```bash
sphinx-build -b html . _build/html
```
Or leverage the make file and run:
```bash
make html
```
If this doesn't work, try this command:
```bash
sphinx-build -C -b html . _build/html
```
For speedier builds in multiprocessor environments, run:
```bash
sphinx-build -b html . _build/html -j auto
```
**NOTE**: If you get an error for the spelling extension, like `Extension error: Could not import extension sphinxcontrib.spelling (exception: The 'enchant' C library was not found and maybe needs to be installed. See https://pyenchant.github.io/pyenchant/install.html`, run `brew install enchant`.
4. Modify the source; please focus on the specific change you are contributing. If you also reformat all the code, it will be hard for us to focus on your change.
5. Rebuild the documentation with `sphinx-build -b html . _build/html`. Always ensure that the docs build without errors and that your changes look correct before pushing your changes to remote.
* If you encounter errors that are unclear, run the build in verbose mode with `sphinx-build -vv -b html . _build/html`.
6. Commit your changes to your branch with a clear, scoped commit messages. Bad: "fixed stuff". Good: "Updated ref IDs in all containers topics".
7. Push your changes to remote (`git push origin`) and create a PR from your branch into `master` or the standing release branch (example: `release-2.27.0`). Answer any default questions in the pull request interface.
* See: [pull request guide](https://help.github.com/articles/creating-a-pull-request/)).
8. Pay attention to any automated CI failures reported in the pull request, and stay involved in the conversation.
Updated process documentation can be found here: [Runbook: Authoring a topic for the Neuron documentation](https://quip-amazon.com/e9B9AM7Npb17/Runbook-Authoring-a-topic-for-the-Neuron-documentation).
## Updating the sitemap
If you add or remove a topic, you must recreate the sitemap. To do so:
1. From a shell, `cd` to the root of this repo (`private-aws-neuron-sdk-staging`) on your local machine.
2. Run the following command: `python3 ./_utilities/create_sitemap.py`. This will generate the sitemap as `sitemap.xml` in the root folder of the repo.
3. Rename the `sitemap.xml` file to `sitemap1.xml`.
4. Move the `sitemap1.xml` file to the `/static` folder, copying over the previous version.
5. Delete the generated `sitemap.xml` file from the root (**not** from `/static`) if you did a copy instead of a move.
6. Push a PR with the updated sitemap to remote and request DougEric review/approve it.
## Finding contributions to work on
Looking at the existing issues is a great way to find something to contribute on. As our projects, by default, use the default GitHub issue labels (enhancement/bug/duplicate/help wanted/invalid/question/wontfix), looking at any 'help wanted' issues is a great place to start.
* Or, if you're so inclined, get on DougEric's Christmas card list by fixing broken links, formatting errors, removing stale topics, and fixing spelling/grammar errors.
## Code of Conduct
This project has adopted the [Amazon Open Source Code of Conduct](https://aws.github.io/code-of-conduct).
For more information see the [Code of Conduct FAQ](https://aws.github.io/code-of-conduct-faq) or contact
opensource-codeofconduct@amazon.com with any additional questions or comments.
## Security issue notifications
If you discover a potential security issue in this project we ask that you notify AWS/Amazon Security via our [vulnerability reporting page](http://aws.amazon.com/security/vulnerability-reporting/). Please do **not** create a public github issue.
## Licensing
See the [LICENSE-DOCUMENTATION](./LICENSE-DOCUMENTATION), [LICENSE-SAMPLECODE](./LICENSE-SAMPLECODE) and [LICENSE-SUMMARY-DOCS-SAMPLES](./LICENSE-SUMMARY-DOCS-SAMPLES) files for our project's licensing. We will ask you to confirm the licensing of your contribution.
We may ask you to sign a [Contributor License Agreement (CLA)](http://en.wikipedia.org/wiki/Contributor_License_Agreement) for larger chan
================================================
FILE: Dockerfile
================================================
FROM python:3.10-slim
RUN apt-get update && apt-get install -y --no-install-recommends \
make enchant-2 git pandoc \
&& rm -rf /var/lib/apt/lists/* \
&& pandoc --version
COPY --from=ghcr.io/astral-sh/uv:latest /uv /usr/local/bin/uv
WORKDIR /docs
COPY requirements.txt .
RUN uv pip install --system -r requirements.txt --extra-index-url=https://pypi.org/simple
ENTRYPOINT ["/bin/bash"]
================================================
FILE: LICENSE-DOCUMENTATION
================================================
*** Documentation:
Creative Commons Attribution-ShareAlike 4.0 International Public License
By exercising the Licensed Rights (defined below), You accept and agree to be bound by the terms and conditions of this Creative Commons Attribution-ShareAlike 4.0 International Public License ("Public License"). To the extent this Public License may be interpreted as a contract, You are granted the Licensed Rights in consideration of Your acceptance of these terms and conditions, and the Licensor grants You such rights in consideration of benefits the Licensor receives from making the Licensed Material available under these terms and conditions.
Section 1 – Definitions.
a. Adapted Material means material subject to Copyright and Similar Rights that is derived from or based upon the Licensed Material and in which the Licensed Material is translated, altered, arranged, transformed, or otherwise modified in a manner requiring permission under the Copyright and Similar Rights held by the Licensor. For purposes of this Public License, where the Licensed Material is a musical work, performance, or sound recording, Adapted Material is always produced where the Licensed Material is synched in timed relation with a moving image.
b. Adapter's License means the license You apply to Your Copyright and Similar Rights in Your contributions to Adapted Material in accordance with the terms and conditions of this Public License.
c. BY-SA Compatible License means a license listed at creativecommons.org/compatiblelicenses, approved by Creative Commons as essentially the equivalent of this Public License.
d. Copyright and Similar Rights means copyright and/or similar rights closely related to copyright including, without limitation, performance, broadcast, sound recording, and Sui Generis Database Rights, without regard to how the rights are labeled or categorized. For purposes of this Public License, the rights specified in Section 2(b)(1)-(2) are not Copyright and Similar Rights.
e. Effective Technological Measures means those measures that, in the absence of proper authority, may not be circumvented under laws fulfilling obligations under Article 11 of the WIPO Copyright Treaty adopted on December 20, 1996, and/or similar international agreements.
f. Exceptions and Limitations means fair use, fair dealing, and/or any other exception or limitation to Copyright and Similar Rights that applies to Your use of the Licensed Material.
g. License Elements means the license attributes listed in the name of a Creative Commons Public License. The License Elements of this Public License are Attribution and ShareAlike.
h. Licensed Material means the artistic or literary work, database, or other material to which the Licensor applied this Public License.
i. Licensed Rights means the rights granted to You subject to the terms and conditions of this Public License, which are limited to all Copyright and Similar Rights that apply to Your use of the Licensed Material and that the Licensor has authority to license.
j. Licensor means the individual(s) or entity(ies) granting rights under this Public License.
k. Share means to provide material to the public by any means or process that requires permission under the Licensed Rights, such as reproduction, public display, public performance, distribution, dissemination, communication, or importation, and to make material available to the public including in ways that members of the public may access the material from a place and at a time individually chosen by them.
l. Sui Generis Database Rights means rights other than copyright resulting from Directive 96/9/EC of the European Parliament and of the Council of 11 March 1996 on the legal protection of databases, as amended and/or succeeded, as well as other essentially equivalent rights anywhere in the world.
m. You means the individual or entity exercising the Licensed Rights under this Public License. Your has a corresponding meaning.
Section 2 – Scope.
a. License grant.
1. Subject to the terms and conditions of this Public License, the Licensor hereby grants You a worldwide, royalty-free, non-sublicensable, non-exclusive, irrevocable license to exercise the Licensed Rights in the Licensed Material to:
A. reproduce and Share the Licensed Material, in whole or in part; and
B. produce, reproduce, and Share Adapted Material.
2. Exceptions and Limitations. For the avoidance of doubt, where Exceptions and Limitations apply to Your use, this Public License does not apply, and You do not need to comply with its terms and conditions.
3. Term. The term of this Public License is specified in Section 6(a).
4. Media and formats; technical modifications allowed. The Licensor authorizes You to exercise the Licensed Rights in all media and formats whether now known or hereafter created, and to make technical modifications necessary to do so. The Licensor waives and/or agrees not to assert any right or authority to forbid You from making technical modifications necessary to exercise the Licensed Rights, including technical modifications necessary to circumvent Effective Technological Measures. For purposes of this Public License, simply making modifications authorized by this Section 2(a)(4) never produces Adapted Material.
5. Downstream recipients.
A. Offer from the Licensor – Licensed Material. Every recipient of the Licensed Material automatically receives an offer from the Licensor to exercise the Licensed Rights under the terms and conditions of this Public License.
B. Additional offer from the Licensor – Adapted Material. Every recipient of Adapted Material from You automatically receives an offer from the Licensor to exercise the Licensed Rights in the Adapted Material under the conditions of the Adapter’s License You apply.
C. No downstream restrictions. You may not offer or impose any additional or different terms or conditions on, or apply any Effective Technological Measures to, the Licensed Material if doing so restricts exercise of the Licensed Rights by any recipient of the Licensed Material.
6. No endorsement. Nothing in this Public License constitutes or may be construed as permission to assert or imply that You are, or that Your use of the Licensed Material is, connected with, or sponsored, endorsed, or granted official status by, the Licensor or others designated to receive attribution as provided in Section 3(a)(1)(A)(i).
b. Other rights.
1. Moral rights, such as the right of integrity, are not licensed under this Public License, nor are publicity, privacy, and/or other similar personality rights; however, to the extent possible, the Licensor waives and/or agrees not to assert any such rights held by the Licensor to the limited extent necessary to allow You to exercise the Licensed Rights, but not otherwise.
2. Patent and trademark rights are not licensed under this Public License.
3. To the extent possible, the Licensor waives any right to collect royalties from You for the exercise of the Licensed Rights, whether directly or through a collecting society under any voluntary or waivable statutory or compulsory licensing scheme. In all other cases the Licensor expressly reserves any right to collect such royalties.
Section 3 – License Conditions.
Your exercise of the Licensed Rights is expressly made subject to the following conditions.
a. Attribution.
1. If You Share the Licensed Material (including in modified form), You must:
A. retain the following if it is supplied by the Licensor with the Licensed Material:
i. identification of the creator(s) of the Licensed Material and any others designated to receive attribution, in any reasonable manner requested by the Licensor (including by pseudonym if designated);
ii. a copyright notice;
iii. a notice that refers to this Public License;
iv. a notice that refers to the disclaimer of warranties;
v. a URI or hyperlink to the Licensed Material to the extent reasonably practicable;
B. indicate if You modified the Licensed Material and retain an indication of any previous modifications; and
C. indicate the Licensed Material is licensed under this Public License, and include the text of, or the URI or hyperlink to, this Public License.
2. You may satisfy the conditions in Section 3(a)(1) in any reasonable manner based on the medium, means, and context in which You Share the Licensed Material. For example, it may be reasonable to satisfy the conditions by providing a URI or hyperlink to a resource that includes the required information.
3. If requested by the Licensor, You must remove any of the information required by Section 3(a)(1)(A) to the extent reasonably practicable.
b. ShareAlike.In addition to the conditions in Section 3(a), if You Share Adapted Material You produce, the following conditions also apply.
1. The Adapter’s License You apply must be a Creative Commons license with the same License Elements, this version or later, or a BY-SA Compatible License.
2. You must include the text of, or the URI or hyperlink to, the Adapter's License You apply. You may satisfy this condition in any reasonable manner based on the medium, means, and context in which You Share Adapted Material.
3. You may not offer or impose any additional or different terms or conditions on, or apply any Effective Technological Measures to, Adapted Material that restrict exercise of the rights granted under the Adapter's License You apply.
Section 4 – Sui Generis Database Rights.
Where the Licensed Rights include Sui Generis Database Rights that apply to Your use of the Licensed Material:
a. for the avoidance of doubt, Section 2(a)(1) grants You the right to extract, reuse, reproduce, and Share all or a substantial portion of the contents of the database;
b. if You include all or a substantial portion of the database contents in a database in which You have Sui Generis Database Rights, then the database in which You have Sui Generis Database Rights (but not its individual contents) is Adapted Material, including for purposes of Section 3(b); and
c. You must comply with the conditions in Section 3(a) if You Share all or a substantial portion of the contents of the database.
For the avoidance of doubt, this Section 4 supplements and does not replace Your obligations under this Public License where the Licensed Rights include other Copyright and Similar Rights.
Section 5 – Disclaimer of Warranties and Limitation of Liability.
a. Unless otherwise separately undertaken by the Licensor, to the extent possible, the Licensor offers the Licensed Material as-is and as-available, and makes no representations or warranties of any kind concerning the Licensed Material, whether express, implied, statutory, or other. This includes, without limitation, warranties of title, merchantability, fitness for a particular purpose, non-infringement, absence of latent or other defects, accuracy, or the presence or absence of errors, whether or not known or discoverable. Where disclaimers of warranties are not allowed in full or in part, this disclaimer may not apply to You.
b. To the extent possible, in no event will the Licensor be liable to You on any legal theory (including, without limitation, negligence) or otherwise for any direct, special, indirect, incidental, consequential, punitive, exemplary, or other losses, costs, expenses, or damages arising out of this Public License or use of the Licensed Material, even if the Licensor has been advised of the possibility of such losses, costs, expenses, or damages. Where a limitation of liability is not allowed in full or in part, this limitation may not apply to You.
c. The disclaimer of warranties and limitation of liability provided above shall be interpreted in a manner that, to the extent possible, most closely approximates an absolute disclaimer and waiver of all liability.
Section 6 – Term and Termination.
a. This Public License applies for the term of the Copyright and Similar Rights licensed here. However, if You fail to comply with this Public License, then Your rights under this Public License terminate automatically.
b. Where Your right to use the Licensed Material has terminated under Section 6(a), it reinstates:
1. automatically as of the date the violation is cured, provided it is cured within 30 days of Your discovery of the violation; or
2. upon express reinstatement by the Licensor.
c. For the avoidance of doubt, this Section 6(b) does not affect any right the Licensor may have to seek remedies for Your violations of this Public License.
d. For the avoidance of doubt, the Licensor may also offer the Licensed Material under separate terms or conditions or stop distributing the Licensed Material at any time; however, doing so will not terminate this Public License.
e. Sections 1, 5, 6, 7, and 8 survive termination of this Public License.
Section 7 – Other Terms and Conditions.
a. The Licensor shall not be bound by any additional or different terms or conditions communicated by You unless expressly agreed.
b. Any arrangements, understandings, or agreements regarding the Licensed Material not stated herein are separate from and independent of the terms and conditions of this Public License.
Section 8 – Interpretation.
a. For the avoidance of doubt, this Public License does not, and shall not be interpreted to, reduce, limit, restrict, or impose conditions on any use of the Licensed Material that could lawfully be made without permission under this Public License.
b. To the extent possible, if any provision of this Public License is deemed unenforceable, it shall be automatically reformed to the minimum extent necessary to make it enforceable. If the provision cannot be reformed, it shall be severed from this Public License without affecting the enforceability of the remaining terms and conditions.
c. No term or condition of this Public License will be waived and no failure to comply consented to unless expressly agreed to by the Licensor.
d. Nothing in this Public License constitutes or may be interpreted as a limitation upon, or waiver of, any privileges and immunities that apply to the Licensor or You, including from the legal processes of any jurisdiction or authority.
================================================
FILE: LICENSE-SAMPLECODE
================================================
Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
Permission is hereby granted, free of charge, to any person obtaining a copy of this
software and associated documentation files (the "Software"), to deal in the Software
without restriction, including without limitation the rights to use, copy, modify,
merge, publish, distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
================================================
FILE: LICENSE-SUMMARY-DOCS-SAMPLES
================================================
*** Documentation and Sample Code:
Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.
The documentation is made available under the Creative Commons Attribution-ShareAlike 4.0 International License. See the LICENSE file.
The sample code within this documentation is made available under the MIT-0 license. See the LICENSE-SAMPLECODE file.
================================================
FILE: Makefile
================================================
# Minimal makefile for Sphinx documentation
#
# You can set these variables from the command line, and also
# from the environment for the first two.
SPHINXOPTS ?=
SPHINXBUILD ?= sphinx-build
SOURCEDIR = $(CURDIR)
BUILDDIR = _build
# Put it first so that "make" without argument is like "make help".
help:
@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
.PHONY: help Makefile clean
# Catch-all target: route all unknown targets to Sphinx using the new
# "make mode" option. $(O) is meant as a shortcut for $(SPHINXOPTS).
%: Makefile
@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
clean:
-rm -rf $(BUILDDIR)/*
================================================
FILE: README.md
================================================
# AWS Neuron
## Neuron SDK Overview
AWS Neuron is a software development kit (SDK) enabling high-performance deep learning acceleration using AWS Inferentia and Trainium, AWS's custom designed machine learning accelerators. With Neuron, you can develop, profile, and deploy high-performance machine learning workloads on top of accelerated EC2 instances, e.g. Inf1 and Trn1.
Neuron includes a compiler, runtime driver, as well as debug and profiling utilities with a TensorBoard plugin for visualization, and is pre-integrated into popular machine learning frameworks like Pytorch, TensorFlow and MXNet, to provide a seamless machine learning acceleration workflow.
## Neuron SDK’s documentation
For full documentations including user guide, Howtos and Tutorials see [Neuron SDK’s documentation](https://awsdocs-neuron.readthedocs-hosted.com/)
## Support
If none of the github and online resources have an answer to your question, checkout the AWS Neuron [support forum](https://forums.aws.amazon.com/forum.jspa?forumID=355).
================================================
FILE: _backup-setup/neuron-setup/multiframework/multi-framework-ubuntu22-neuron-dlami.rst
================================================
.. _setup-ubuntu22-multi-framework-dlami:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
Get Started with Neuron on Ubuntu 22 with Neuron Multi-Framework DLAMI
======================================================================
You can quickly get started on Ubuntu 22 using the Neuron Deep Learning AMI (DLAMI). Then, start using one of the multiple frameworks or libraries that Neuron SDK supports by
activating the corresponding virtual environment. Each virtual environment comes pre-installed with Neuron libraries needed for you to get started. The Neuron DLAMI supports all Neuron instances (Inf1/Inf2/Trn1/Trn1n/Trn2/Trn3)
and is updated with each Neuron SDK release. To start using the latest version of the Neuron DLAMI, use the following steps:
Step 1: Launch the instance using Neuron DLAMI
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Once you open the `EC2 Console <https://console.aws.amazon.com/ec2>`_, select your desired AWS region and choose "Launch Instance". Under AMI selection select the "Quick Start"
and "Ubuntu", choose the "Deep Learning AMI Neuron (Ubuntu 22.04)"(see screenshot below). Once you have selected the AMI, select the desired Neuron Instance(Inf1/Inf2/Trn1/Trn1n/Trn2/Trn3) ,
configure disk size and other criteria, launch the instance
.. image:: /images/neuron-multi-framework-dlami-quick-start.png
:scale: 20%
:align: center
.. note::
If you are looking to use the Neuron DLAMI in your cloud automation flows , Neuron also supports :ref:`SSM parameters <ssm-parameter-neuron-dlami>` to easily retrieve the latest DLAMI id.
Step 2: Activate the desired virtual environment
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
You can activate one of the virtual environments depending on the library or framework you are interested in:
1. Get the desired virtual environment name for the framework/library by referring to :ref:`the Neuron DLAMI overview <neuron-dlami-multifw-venvs>`.
2. Activate the virtual environment by using:
::
source /opt/<name_of_virtual_environment>/bin/activate
After you have activated the desired virtual environment , you can try out one of the tutorials listed in the corresponding framework or library training and inference section.
================================================
FILE: _backup-setup/neuron-setup/multiframework/multi-framework-ubuntu24-neuron-dlami.rst
================================================
.. _setup-ubuntu24-multi-framework-dlami:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
Get Started with Neuron on Ubuntu 24 with Neuron Multi-Framework DLAMI
======================================================================
You can quickly get started on Ubuntu 24 using the Neuron Deep Learning AMI (DLAMI). Then, start using one of the multiple frameworks or libraries that Neuron SDK supports by
activating the corresponding virtual environment. Each virtual environment comes pre-installed with Neuron libraries needed for you to get started. The Neuron DLAMI supports all Neuron instances (Inf2/Trn1/Trn1n/Trn2/Trn3)
and is updated with each Neuron SDK release. To start using the latest version of the Neuron DLAMI, use the following steps:
Step 1: Launch the instance using Neuron DLAMI
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Once you open the `EC2 Console <https://console.aws.amazon.com/ec2>`_, select your desired AWS region and choose "Launch Instance". Under AMI selection select the "Quick Start"
and "Ubuntu", choose the "Deep Learning AMI Neuron (Ubuntu 24.04)"(see screenshot below). Once you have selected the AMI, select the desired Neuron Instance(Inf2/Trn1/Trn1n/Trn2/Trn3),
configure disk size and other criteria, launch the instance
.. image:: /images/neuron-multi-framework-dlami-U24-quick-start.png
:scale: 20%
:align: center
.. note::
If you are looking to use the Neuron DLAMI in your cloud automation flows , Neuron also supports :ref:`SSM parameters <ssm-parameter-neuron-dlami>` to easily retrieve the latest DLAMI id.
Step 2: Activate the desired virtual environment
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
You can activate one of the virtual environments depending on the library or framework you are interested in:
1. Get the desired virtual environment name for the framework/library by referring to :ref:`the Neuron DLAMI overview <neuron-dlami-multifw-venvs>`.
2. Activate the virtual environment by using:
::
source /opt/<name_of_virtual_environment>/bin/activate
After you have activated the desired virtual environment , you can try out one of the tutorials listed in the corresponding framework or library training and inference section.
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuron/amazon-linux/torch-neuron-al2-base-dlami.rst
================================================
.. _setup-torch-neuron-al2-base-dlami:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuron") Setup on Amazon Linux 2 with DLAMI Base
=======================================================================
.. note::
As of 2.20.0, Neuron Runtime no longer supports AL2. Upgrade to AL2023 following the :ref:`AL2 Migration guide <eos-al2>`
.. contents:: Table of contents
:local:
:depth: 2
.. include:: /setup/install-templates/al2-python.rst
Get Started with Latest Release of PyTorch Neuron (``torch-neuron``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`setup-torch-neuron` for Inference.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instance sizes and pricing see: `Inf1 web page <https://aws.amazon.com/ec2/instance-types/inf1/>`_
* Check for the latest version of the `DLAMI Base AMI <https://aws.amazon.com/releasenotes/aws-deep-learning-ami-base-neuron-amazon-linux-2/>`_ and copy the AMI name that starts with "Deep Learning Base Neuron AMI (Amazon Linux 2) <latest_date>" from "AMI Name:" section
* Search for the copied AMI name in the AMI Search , you should see a matching AMI with the AMI name in Community AMIs. Select the AMI and use it to launch the instance.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=install --framework=pytorch --framework-version=1.13.1 --file=src/helperscripts/n2-manifest.json --os=amazonlinux2 --instance=inf1 --ami=non-dlami --category=driver_runtime_tools
.. include:: /includes/setup/tab-inference-torch-neuron-al2.txt
.. include :: /archive/torch-neuron/setup/pytorch-update-al2.rst
.. include :: /archive/torch-neuron/setup/pytorch-install-prev-al2.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuron/amazon-linux/torch-neuron-al2-pytorch-dlami.rst
================================================
.. _setup-torch-neuron-al2-pytorch-dlami:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuron") Setup on Amazon Linux 2 with Pytorch DLAMI
=========================================================================
.. note::
As of 2.20.0, Neuron Runtime no longer supports AL2. Upgrade to AL2023 following the :ref:`AL2 Migration guide <eos-al2>`
.. contents:: Table of contents
:local:
:depth: 2
.. include:: /setup/install-templates/al2-python.rst
Get Started with Latest Release of PyTorch Neuron (``torch-neuron``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`setup-torch-neuron`.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Inf1 web page <https://aws.amazon.com/ec2/instance-types/inf1/>`_
* Check for the latest version of the `DLAMI Neuron Pytorch 1.13 AMI <https://aws.amazon.com/releasenotes/aws-deep-learning-ami-neuron-pytorch-1-13-amazon-linux-2/>`_ and copy the AMI name that starts with "Deep Learning AMI Neuron PyTorch 1.13 (Amazon Linux 2) <latest_date>" from "AMI Name:" section
* Search for the copied AMI name in the AMI Search , you should see an exact matching AMI with the AMI name in Community AMIs. Select the AMI and use it to launch the instance.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Update Neuron Drivers
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=update --category=driver_runtime_tools --framework=pytorch --framework-version=1.13.0 --file=src/helperscripts/n2-manifest.json --os=amazonlinux2 --instance=inf1
.. dropdown:: Get Started With Pytorch DLAMI
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 98
:end-line: 99
.. card:: Visit PyTorch Neuron(``torch-neuron``) for Inference section
:link: inference-torch-neuron
:link-type: ref
:class-body: sphinx-design-class-title-small
.. card:: Visit PyTorch Neuron section for more
:class-body: sphinx-design-class-body-small
:link: neuron-pytorch
:link-type: ref
.. include:: /archive/torch-neuron/setup/pytorch-update-al2-dlami.rst
.. include:: /archive/torch-neuron/setup/pytorch-install-prev-al2.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuron/amazon-linux/torch-neuron-al2.rst
================================================
.. _setup-torch-neuron-al2:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuron") Setup on Amazon Linux 2
=========================================================
.. note::
As of 2.20.0, Neuron Runtime no longer supports AL2. Upgrade to AL2023 following the :ref:`AL2 Migration guide <eos-al2>`
.. contents:: Table of contents
:local:
:depth: 2
.. include:: /setup/install-templates/al2-python.rst
Get Started with Latest Release of PyTorch Neuron (``torch-neuron``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`setup-torch-neuron` for Inference.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console, please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Inf1 web page <https://aws.amazon.com/ec2/instance-types/inf1/>`_
* Select Amazon Linux 2 AMI(HVM) - Kernel 5.10
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=install --framework=pytorch --framework-version=1.13.1 --file=src/helperscripts/n2-manifest.json --os=amazonlinux2 --instance=inf1 --ami=non-dlami --category=driver_runtime_tools
.. include:: /includes/setup/tab-inference-torch-neuron-al2.txt
.. include :: /archive/torch-neuron/setup/pytorch-update-al2.rst
.. include :: /archive/torch-neuron/setup/pytorch-install-prev-al2.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuron/amazon-linux/torch-neuron-al2023.rst
================================================
.. _setup-torch-neuron-al2023:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuron") Setup on Amazon Linux 2023
===========================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuron``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`setup-torch-neuron` for Inference.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console, please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Inf1 web page <https://aws.amazon.com/ec2/instance-types/inf1/>`_
* Select Amazon Linux 2023 AMI
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=install --framework=pytorch --framework-version=1.13.1 --file=src/helperscripts/n2-manifest.json --os=amazonlinux2 --instance=inf1 --ami=non-dlami --category=driver_runtime_tools
.. include:: /includes/setup/tab-inference-torch-neuron-al2023.txt
.. include:: /archive/torch-neuron/setup/pytorch-update-al2023.rst
.. include:: /archive/torch-neuron/setup/pytorch-install-prev-al2023.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuron/ubuntu/torch-neuron-ubuntu20-base-dlami.rst
================================================
.. _setup-torch-neuron-u20-base-dlami:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuron") Setup on Ubuntu 20 with DLAMI Base
==================================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuron``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`setup-torch-neuron` for Inference.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instance sizes and pricing see: `Inf1 web page <https://aws.amazon.com/ec2/instance-types/inf1/>`_
* Check for the latest version of the `DLAMI Base AMI <https://aws.amazon.com/releasenotes/aws-deep-learning-ami-base-neuron-ubuntu-20-04/>`_ and copy the AMI name that starts with "Deep Learning Base Neuron AMI (Ubuntu 20.04) <latest_date>" from "AMI Name:" section
* Search for the copied AMI name in the AMI Search , you should see a matching AMI with the AMI name in Community AMIs. Select the AMI and use it to launch the instance.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=install --framework=pytorch --framework-version=1.13.1 --file=src/helperscripts/n2-manifest.json --os=ubuntu20 --instance=inf1 --ami=non-dlami --category=driver_runtime_tools
.. include:: /includes/setup/tab-inference-torch-neuron-u20.txt
.. include:: /archive/torch-neuron/setup/pytorch-update-u20.rst
.. include:: /archive/torch-neuron/setup/pytorch-install-prev-u20.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuron/ubuntu/torch-neuron-ubuntu20-pytorch-dlami.rst
================================================
.. _setup-torch-neuron-u20-pytorch-dlami:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuron") Setup on Ubuntu 20 with Pytorch DLAMI
=====================================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuron``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`setup-torch-neuron`.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Inf1 web page <https://aws.amazon.com/ec2/instance-types/inf1/>`_
* Check for the latest version of the `DLAMI Neuron Pytorch 1.13 AMI <https://aws.amazon.com/releasenotes/aws-deep-learning-ami-neuron-pytorch-1-13-ubuntu-20-04/>`_ and copy the AMI name that starts with "Deep Learning AMI Neuron PyTorch 1.13 (Ubuntu 20.04) <latest_date>" from "AMI Name:" section
* Search for the copied AMI name in the AMI Search , you should see an exact matching AMI with the AMI name in Community AMIs. Select the AMI and use it to launch the instance.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Update Neuron Drivers
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=update --category=driver_runtime_tools --framework=pytorch --framework-version=1.13.0 --file=src/helperscripts/n2-manifest.json --os=ubuntu20 --instance=inf1
.. dropdown:: Get Started With Pytorch DLAMI
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 101
:end-line: 102
.. card:: PyTorch Neuron(``torch-neuron``) for Inference
:link: inference-torch-neuron
:link-type: ref
:class-body: sphinx-design-class-title-small
.. card:: Visit PyTorch Neuron section for more
:class-body: sphinx-design-class-body-small
:link: neuron-pytorch
:link-type: ref
.. include:: /archive/torch-neuron/setup/pytorch-update-u20-dlami.rst
.. include:: /archive/torch-neuron/setup/pytorch-install-prev-u20.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuron/ubuntu/torch-neuron-ubuntu20.rst
================================================
.. _setup-torch-neuron-u20:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuron") Setup on Ubuntu 20
====================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuron``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`setup-torch-neuron` for Inference.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Inf1 web page <https://aws.amazon.com/ec2/instance-types/inf1/>`_
* Select Ubuntu Server 20 AMI
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=install --framework=pytorch --framework-version=1.13.1 --file=src/helperscripts/n2-manifest.json --os=ubuntu20 --instance=inf1 --ami=non-dlami --category=driver_runtime_tools
.. include:: /includes/setup/tab-inference-torch-neuron-u20.txt
.. include:: /archive/torch-neuron/setup/pytorch-update-u20.rst
.. include:: /archive/torch-neuron/setup/pytorch-install-prev-u20.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuron/ubuntu/torch-neuron-ubuntu22.rst
================================================
.. _setup-torch-neuron-u22:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuron") Setup on Ubuntu 22
=====================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuron``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`setup-torch-neuron` for Inference.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Inf1 web page <https://aws.amazon.com/ec2/instance-types/inf1/>`_
* Select Ubuntu Server 22 AMI
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=install --framework=pytorch --framework-version=1.13.1 --file=src/helperscripts/n2-manifest.json --os=ubuntu20 --instance=inf1 --ami=non-dlami --category=driver_runtime_tools
.. include:: /includes/setup/tab-inference-torch-neuron-u22.txt
.. include:: /archive/torch-neuron/setup/pytorch-update-u22.rst
.. include:: /archive/torch-neuron/setup/pytorch-install-prev-u22.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/amazon-linux/torch-neuronx-al2-base-dlami.rst
================================================
.. _setup-torch-neuronx-al2-base-dlami:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Amazon Linux 2 with DLAMI Base
=========================================================================
.. note::
As of 2.20.0, Neuron Runtime no longer supports AL2. Upgrade to AL2023 following the :ref:`AL2 Migration guide <eos-al2>`
.. contents:: Table of contents
:local:
:depth: 2
.. include:: /setup/install-templates/al2-python.rst
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instance sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Check for the latest version of the `DLAMI Base AMI <https://aws.amazon.com/releasenotes/aws-deep-learning-ami-base-neuron-amazon-linux-2/>`_ and copy the AMI name that starts with "Deep Learning Base Neuron AMI (Amazon Linux 2) <latest_date>" from "AMI Name:" section
* Search for the copied AMI name in the AMI Search , you should see a matching AMI with the AMI name in Community AMIs. Select the AMI and use it to launch the instance.
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 2
:end-line: 3
.. include:: /includes/setup/tab-inference-torch-neuronx-al2.txt
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-al2.rst
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-install-prev-al2.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/amazon-linux/torch-neuronx-al2-pytorch-dlami.rst
================================================
.. _setup-torch-neuronx-al2-dlami-pytorch:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Amazon Linux 2 with DLAMI Pytorch
===========================================================================
.. note::
As of 2.20.0, Neuron Runtime no longer supports AL2. Upgrade to AL2023 following the :ref:`AL2 Migration guide <eos-al2>`
.. contents:: Table of contents
:local:
:depth: 2
.. include:: /setup/install-templates/al2-python.rst
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Check for the latest version of the `DLAMI Neuron Pytorch 1.13 AMI <https://aws.amazon.com/releasenotes/aws-deep-learning-ami-neuron-pytorch-1-13-amazon-linux-2/>`_ and copy the AMI name that starts with "Deep Learning AMI Neuron PyTorch 1.13 (Amazon Linux 2) <latest_date>" from "AMI Name:" section
* Search for the copied AMI name in the AMI Search , you should see an exact matching AMI with the AMI name in Community AMIs. Select the AMI and use it to launch the instance.
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Update Neuron Drivers
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=update --category=driver_runtime_tools --framework=pytorch --framework-version=2.9.0 --file=src/helperscripts/n2-manifest.json --os=amazonlinux2 --instance=trn1
.. dropdown:: Get Started With Pytorch DLAMI
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 50
:end-line: 51
.. card:: Visit PyTorch Neuron(``torch-neuronx``) for Inference section
:link: inference-torch-neuronx
:link-type: ref
:class-body: sphinx-design-class-title-small
.. card:: Visit PyTorch Neuron(``torch-neuronx``) for Training section
:link: training-torch-neuronx
:link-type: ref
:class-body: sphinx-design-class-title-small
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-al2-dlami.rst
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-install-prev-al2.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/amazon-linux/torch-neuronx-al2.rst
================================================
.. _setup-torch-neuronx-al2:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Amazon Linux 2
=========================================================
.. note::
As of 2.20.0, Neuron Runtime no longer supports AL2. Upgrade to AL2023 following the :ref:`AL2 Migration guide <eos-al2>`
.. contents:: Table of contents
:local:
:depth: 2
.. include:: /setup/install-templates/al2-python.rst
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console, please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Select Amazon Linux 2 AMI(HVM) - Kernel 5.10
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 2
:end-line: 3
.. include:: /includes/setup/tab-inference-torch-neuronx-al2.txt
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-al2.rst
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-install-prev-al2.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/amazon-linux/torch-neuronx-al2023.rst
================================================
.. _setup-torch-neuronx-al2023:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Amazon Linux 2023
============================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console, please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Select Amazon Linux 2023 AMI
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 239
:end-line: 240
.. include:: /includes/setup/tab-inference-torch-neuronx-al2023.txt
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-al2023.rst
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-install-prev-al2023.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu20-base-dlami.rst
================================================
.. _setup-torch-neuronx-ubuntu20-base-dlami:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Ubuntu 20 with DLAMI Base
====================================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instance sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Check for the latest version of the `DLAMI Base AMI <https://aws.amazon.com/releasenotes/aws-deep-learning-ami-base-neuron-ubuntu-20-04/>`_ and copy the AMI name that starts with "Deep Learning Base Neuron AMI (Ubuntu 20.04) <latest_date>" from "AMI Name:" section
* Search for the copied AMI name in the AMI Search , you should see a matching AMI with the AMI name in Community AMIs. Select the AMI and use it to launch the instance.
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 5
:end-line: 6
.. include:: /includes/setup/tab-inference-torch-neuronx-u20.txt
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-u20.rst
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-install-prev-u20.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu20-pytorch-dlami.rst
================================================
.. _setup-torch-neuronx-ubuntu20-dlami-pytorch:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Ubuntu 20 with DLAMI Pytorch
======================================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Check for the latest version of the `DLAMI Neuron Pytorch 1.13 AMI <https://aws.amazon.com/releasenotes/aws-deep-learning-ami-neuron-pytorch-1-13-ubuntu-20-04/>`_ and copy the AMI name that starts with "Deep Learning AMI Neuron PyTorch 1.13 (Ubuntu 20.04) <latest_date>" from "AMI Name:" section
* Search for the copied AMI name in the AMI Search , you should see an exact matching AMI with the AMI name in Community AMIs. Select the AMI and use it to launch the instance.
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Update Neuron Drivers
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
.. program-output:: python3 src/helperscripts/n2-helper.py --install-type=update --category=driver_runtime_tools --framework=pytorch --framework-version=2.9.0 --file=src/helperscripts/n2-manifest.json --os=ubuntu20 --instance=trn1
.. dropdown:: Get Started With Pytorch DLAMI
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 53
:end-line: 54
.. card:: Visit PyTorch Neuron(``torch-neuronx``) for Inference section
:link: inference-torch-neuronx
:link-type: ref
:class-body: sphinx-design-class-title-small
.. card:: Visit PyTorch Neuron(``torch-neuronx``) for Training section
:link: training-torch-neuronx
:link-type: ref
:class-body: sphinx-design-class-title-small
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-u20-dlami.rst
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-install-prev-u20.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu20.rst
================================================
.. _setup-torch-neuronx-ubuntu20:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:width: 100%
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Ubuntu 20
===================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. include:: /setup/install-templates/trn1-ga-warning.txt
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console. please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Select Ubuntu Server 20 AMI
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 5
:end-line: 6
.. include:: /includes/setup/tab-inference-torch-neuronx-u20.txt
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-u20.rst
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-install-prev-u20.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu22.rst
================================================
.. _setup-torch-neuronx-ubuntu22:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Ubuntu 22
=====================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. include:: /setup/install-templates/trn1-ga-warning.txt
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console, please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Select Ubuntu Server 22 AMI
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 242
:end-line: 243
.. include:: /includes/setup/tab-inference-torch-neuronx-u22.txt
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-u22.rst
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-install-prev-u22.rst
================================================
FILE: _backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu24.rst
================================================
.. _setup-torch-neuronx-ubuntu24:
.. card:: Select a Different Framework or Platform for Setup
:link: setup-guide-index
:link-type: ref
:class-body: sphinx-design-class-title-small
PyTorch Neuron ("torch-neuronx") Setup on Ubuntu 24
=====================================================
.. contents:: Table of contents
:local:
:depth: 2
Get Started with Latest Release of PyTorch Neuron (``torch-neuronx``)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This section provide links that will assist you to quickly start with a fresh installation of :ref:`pytorch-neuronx-main` for both Inference and Training.
.. include:: /setup/install-templates/trn1-ga-warning.txt
.. dropdown:: Launch the Instance
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
* Please follow the instructions at `launch an Amazon EC2 Instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/EC2_GetStarted.html#ec2-launch-instance>`_ to launch an instance. When choosing the instance type at the EC2 console, please make sure to select the correct instance type.
* To get more information about instances sizes and pricing see: `Trn1 web page <https://aws.amazon.com/ec2/instance-types/trn1/>`_, `Inf2 web page <https://aws.amazon.com/ec2/instance-types/inf2/>`_
* Select Ubuntu Server 24 AMI
* When launching a Trn1, please adjust your primary EBS volume size to a minimum of 512GB.
* After launching the instance, follow the instructions in `Connect to your instance <https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AccessingInstancesLinux.html>`_ to connect to the instance
.. dropdown:: Install Drivers and Tools
:class-title: sphinx-design-class-title-small
:class-body: sphinx-design-class-body-small
:animate: fade-in
.. include:: /src/helperscripts/installationScripts/python_instructions.txt
:start-line: 299
:end-line: 300
.. include:: /includes/setup/tab-inference-torch-neuronx-u24.txt
.. include:: /frameworks/torch/torch-neuronx/setup/pytorch-update-u24.rst
================================================
FILE: _content-types/conceptual-deep-dive.rst
================================================
.. meta::
:description: {short description here}
:date_updated: {planned date of publication here}
.. _{RST page ref string here}:
================================================================================
Deep dive: {concept/practice/technique name; use sentence-case, not title case!}
================================================================================
.. {SEO-friendly intro paragraph, no more than 3 sentences total.}
This topic explores {subjects} in depth and discusses the technical details of it from the perspective of an AWS Neuron expert. Some experience in {related subjects here} is required to understand it in full.
What you should know before reading
-----------------------------------
.. {If there is anything the reader should know before diving into this material, note it here and provide any supporting links. This also helps LLMs training on this content have greater technical context for this subject.}
Before you start, you must be familiar with the following:
- **Concept 1:** {Brief description. Link to a related topic if necessary.}
- **Concept 2:** {Brief description. Link to a related topic if necessary.}
Overview
---------
.. {Your first section, which should cover the subject from the title at a high level. If appropriate, note when this concept is applicable in Neuron components and developer workflows. Starting off with a diagram can help illustrate the concept.}
PARAGRAPH 1
PARAGRAPH 2
.. image:: images/diagram-name.png
:alt: {Alt text for diagram}
:align: center
{Section 1 Title}
-----------------
.. {Each section should build on top of what was discussed in the previous sections. If a new concept is introduced that wasn't discussed previously, link to a topic that covers it. You can add subsections within this section if it helps to break it up more and clarify the content, but do not go more than 1-2 levels deep.}
PARAGRAPH 1
PARAGRAPH 2
.. code-block:: python
# Code example if applicable
def example_function():
pass
{Section 2 Title}
-----------------
.. {Each section should build on top of what was discussed in the previous sections. If a new concept is introduced that wasn't discussed previously, link to a topic that covers it. You can add subsections within this section if it helps to break it up more and clarify the content, but do not go more than 1-2 levels deep.}
PARAGRAPH 1
PARAGRAPH 2
.. code-block:: python
# Code example if applicable
def example_function():
pass
.. {Add more sections as appropriate to logically break up the content. Each section should be focused on a specific aspect of the concept.}
{optional}Related Concepts
----------------
* :ref:`link-reference-name` - {description}
* :ref:`link-reference-name` - {description}
{optional}Further Reading
---------------
.. toctree::
:maxdepth: 1
* `External Link <URL>`_ - {description}
* :doc:`/path/to/internal/doc` - {description}
.. (Note to both the writer and any AI incorporating this template: The content below is provided as a resource and should not be included as-is in any final document created using this template as a basis.)
.. note::
.. Additional implementation details or important considerations can be added as admonitions.
.. warning::
.. Critical information or potential pitfalls can be highlighted using warning admonitions.
================================================
FILE: _content-types/model-card.rst
================================================
.. _unique-ref-id-here:
.. meta::
:description: AWS Neuron SDK model card for {Model Name}, version {version}. Overview, intended use, training data, performance, limitations, ethical considerations, and citations.
:date-modified: 2026-10-03
Model Card: {Model Name}
=======================
.. contents:: Table of Contents
:depth: 1
:local:
Model overview
--------------
:Model name: {name}
:Version: {version}
:Organization: {organization}
:License: {license}
:Last updated: {date}
.. warning::
{Important warnings or critical limitations}
Quickstart
----------
.. code-block:: python
# Example usage code
from model import Model
model = Model.from_pretrained("model_name")
output = model.generate("Your input text")
Model details
-------------
Architecture
^^^^^^^^^^^^
- Base architecture: {architecture}
- Number of parameters: {parameter_count}
- Model dimensions: {model_dimensions}
- Training objective: {training_objective}
Hardware requirements
^^^^^^^^^^^^^^^^^^^^^
- Minimum RAM: {min_ram}
- Recommended GPU: {gpu_specs}
- Disk space: {disk_space}
Intended Use
-----------
Primary uses
^^^^^^^^^^^^
* {use_case_1}
* {use_case_2}
* {use_case_3}
Out-of-Scope uses
^^^^^^^^^^^^^^^^^
* {prohibited_use_1}
* {prohibited_use_2}
Training data
------------
Datasets
^^^^^^^^
.. list-table::
:header-rows: 1
* - Dataset Name
- Size
- Description
* - {dataset1}
- {size1}
- {description1}
* - {dataset2}
- {size2}
- {description2}
Training procedure
^^^^^^^^^^^^^^^^^^
* Training hardware: {hardware_details}
* Training time: {duration}
* Training cost: {cost_estimate}
* Carbon footprint: {carbon_impact}
Performance and limitations
---------------------------
Benchmarks
^^^^^^^^^
.. list-table::
:header-rows: 1
* - Benchmark
- Score
- Details
* - {benchmark1}
- {score1}
- {details1}
* - {benchmark2}
- {score2}
- {details2}
Known limitations
^^^^^^^^^^^^^^^^^
* {limitation_1}
* {limitation_2}
Bias and fairness
^^^^^^^^^^^^^^^^^
* {bias_consideration_1}
* {bias_consideration_2}
Ethical considerations
----------------------
Potential risks
^^^^^^^^^^^^^^^
* {risk_1}
* {risk_2}
Mitigation strategies
^^^^^^^^^^^^^^^^^^^^^
* {strategy_1}
* {strategy_2}
Model details and notes
----------------------
{Provide detailed information about the model, its training, evaluation, and any other relevant aspects. Create the sections as needed.}
{Section 1 title}
^^^^^^^^^^^^^^^^^
{Details for section 1.}
{Section 2 title}
^^^^^^^^^^^^^^^^^
{Details for section 2.}
{. . .}
Citations
---------
.. code-block:: bibtex
@article{model_paper,
title={},
author={},
journal={},
year={}
}
Version history
---------------
.. list-table::
:header-rows: 1
* - Version
- Date
- Changes
* - {version1}
- {date1}
- {changes1}
* - {version2}
- {date2}
- {changes2}
Contact
-------
:Documentation Issues: {link_to_issues}
:Support: {support_contact}
:Website: {website_url}
================================================
FILE: _content-types/procedural-how-to.rst
================================================
.. meta::
:description: {short description here}
:date_updated: {planned date of publication here}
.. _{RST page ref string here}:
========================================================================
How to {verb phrase with specific features or models that will be used}
========================================================================
Task overview
-------------
.. {SEO-friendly intro paragraph, no more than 3 sentences total.}
This topic discusses how to {description of task or process here} using the AWS Neuron SDK. {Short description of what the task will accomplish.}
Prerequisites
-------------
- **Prerequisite 1:** Description. Link to a related topic if necessary.
- **Prerequisite 2:** Description. Link to a related topic if necessary.
Instructions
------------
**1:** {First step; start with verb/action}
.. {Describe what the user will do in this step, starting with a verb. If applicable, include any commands or code examples that illustrate the step.}
.. code-block:: bash
# Command or code example
command --flag value
.. {Additional detail if needed.}
.. note::
.. {Optional; important information or caveats about this step}
**2:** {Second step; start with verb/action}
.. .. {Describe what the user will do in this step, starting with a verb. If applicable, include any commands or code examples that illustrate the step.}
.. code-block:: python
# Code example if applicable
def example():
pass
.. {Additional detail if needed.}
.. note::
.. {Optional; important information or caveats about this step}
.. **{More discrete steps as needed, following the same pattern as above.}**
**N:** {Last step; start with verb/action}
.. {Final step instructions}
Confirm your work
-----------------
To confirm you have successfully completed this task, {how to verify the task was done correctly}:
.. {Provide them with a way to know they’ve done everything correctly. This could be a screenshot, command-line output, a tool to launch, or specific settings to check.}
.. code-block:: bash
# Verification command if applicable
verify-command --check
Common issues
-------------
Uh oh! Did you encounter an error or other issue while working through this task? Here are some commonly encountered issues and how to address them.
.. rubric:: {Problem 1}
- **Possible solution**: {detailed solution}
.. rubric:: {Problem 2}
- **Possible solution**: {detailed solution}
Related information
-------------------
.. toctree::
:maxdepth: 1
* `External Link <URL>`_ - {description}
* :doc:`/path/to/internal/doc` - {description}
================================================
FILE: _content-types/procedural-tutorial.ipynb
================================================
{
"cells": [
{
"cell_type": "raw",
"metadata": {
"raw_mimetype": "text/restructuredtext",
"vscode": {
"languageId": "raw"
}
},
"source": [
".. meta::\n",
" :description: {SEO-friendly short description of the tutorial. Include 'Neuron' and any keywords such as the language mode and framework.}\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Tutorial: {title starting with verb}\n",
"\n",
"This tutorial guides you through using the AWS Neuron SDK to {description of what the reader will accomplish in this tutorial, using a specific component or framework}.\n",
"\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Overview\n",
"\n",
"{Briefly summarize the purpose and outcome of this end-to-end tutorial}.\n",
"{State what users will learn or achieve by completing the tutorial}."
]
},
{
"cell_type": "raw",
"metadata": {
"raw_mimetype": "text/restructuredtext",
"vscode": {
"languageId": "raw"
}
},
"source": [
".. contents:: Table of contents\n",
" :local:\n",
" :depth: 2"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Before you start\n",
"\n",
"To successfully complete this tutorial, you must have completed the following steps in advance:\n\n",
"- Downloaded and installed the [AWS Neuron SDK](https://awsdocs-neuron.readthedocs-hosted.com/en/latest/setup/index.html) for {component}.\n",
"- {prerequisite 2 description here. If the user must read a topic in advance or perform any complex preparations, provide a link to a topic or download}\n",
"- {prerequisite 3 description here}\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Setup\n",
"\n",
"{Describe any initial local setup required before starting the tutorial.}\n",
"{Include any code-specific installation, configuration, or environment setup steps.}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# Example setup command (Remove these comments and add the CLI commands, env variable declarations, or other operations for the user to prepare their environment.)\n",
"# pip install package_name"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Tutorial steps"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Step 1: {Title, starting with an infinitive verb like 'Load...', 'Create...', etc.}\n",
"\n",
"{Describe the first main step. Provide code, commands, or configuration as needed.}\n",
"\n",
"{Optional} {Add any important notes, caveats, or warnings for this step.}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"print(\"Code goes here!\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Step 2: {Title, starting with an infinitive verb like 'Load...', 'Create...', etc.}\n",
"\n",
"{Describe the second main step.}"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"print(\"Code goes here!\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Step 3: {Title, starting with an infinitive verb like 'Load...', 'Create...', etc.}\n",
"\n",
"Describe the third main step."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"print(\"Code goes here!\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Step N: {Title, starting with an infinitive verb like 'Load...', 'Create...', etc.}\n",
"\n",
"Describe the last main step."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"print(\"Code goes here!\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\nCode completed. Now, let's run it..."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Run the code\n",
"\n",
"To run this code, {action to take to run the code}:\n",
"Include commands, expected outputs, or checks to perform."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"# Example verification command\n",
"# python foo.py\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\n",
"If your code works, you will see output like this:\n\n",
"```\n",
"Loading glorp inhume logic...Done!\n",
"Configuring extubation channel instances...Done!\n\n",
"1111 | 2222 | 3333\n",
"4444 | 5555 | 6666\n\n",
"Average glorps inhumed and extubated: 420\n",
"Time to max glorp: 8 seconds\n",
"```\n\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"\nCongratulations! You now know how to {goal of tutorial}. If your code did not run or did not produce similar results, see the [Common issues](#Common issues) section below."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Common issues\n",
"\n",
"Here are some common errors and mistakes you can make when developing code using the approach in this tutorial, and how you may be able to address them:\n\n",
"- {describe error, symptoms, and possible solution}\n",
"- {describe error, symptoms, and possible solution}"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## (Optional) Next steps\n",
"\n",
"{Suggest what users might want to do next after completing the tutorial.\n",
"Link to related topics or advanced guides.}"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Related topics\n",
"\n",
"- [Related topic 1](link_here)\n",
"- [Related topic 2](link_here)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.8.5"
}
},
"nbformat": 4,
"nbformat_minor": 4
}
================================================
FILE: _content-types/reference-kernel-api.rst
================================================
.. meta::
:description: API reference for the {kernel-name} kernel included in the NKI Library .
:date-modified: MM/DD/YYYY
.. currentmodule:: {kernel namespace}.{kernel module path}
RMSNorm-Quant Kernel API Reference
==================================
This topic provides the API reference for the ``{kernel name}`` kernel. The kernel performs optional RMS normalization followed by quantization to ``fp8``.
The kernel supports:
* {feature 1}
* {feature 2}
* {feature 3}
* ... {more features as needed}
Background
-----------
The ``{kernel}`` kernel ... {description of kernel functionality based in sources}
For detailed information about the mathematical operations and implementation details, refer to the :doc:`{kernel name} Kernel Design Specification </nki/library/specs/{kernel-spec-doc-file-link}>`.
API Reference
--------------
{kernel argument class name}
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
.. py:class:: {kernel argument class name}
{kernel name} Kernel arguments.
.. py:attribute:: {attribute-1}
:type: {attribute-1-type}
{description from docstring}
.. py:attribute:: {attribute-1}
:type: {attribute-1-type}
{description from docstring}
{more attributes as needed}
.. py:method:: {method syntax} -> {return type}
{description from docstring}
.. py:method:: {method syntax} -> {return type}
{description from docstring}
**Raises**:
* **{exception-1}** – {when exception is raised}
* **{exception-1}** – {when exception is raised}
{kernel API function name in code}
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
.. py:function:: rmsnorm_quant_kernel(hidden: nt.tensor, ln_w: nt.tensor, kargs: RmsNormQuantKernelArgs)
{definition of method used to instantiate or invoke kernel here, from source docstrings}
{params and types with descriptions from source docstrings}
Implementation Details
-----------------------
The kernel implementation includes several key optimizations:
1. **{optimization-or-feature}**: {description}
2. **{optimization-or-feature}**: {description}
3. **{optimization-or-feature}**: {description}
Example
--------
The following is a simple example of how to use the ``{kernel}`` kernel:
.. code-block:: python
# Code here -- need usage example in pedagogical style.
See Also
--------
* :doc:`{kernel} </nki/library/specs/{link-to-kernel-spec}>`
================================================
FILE: _content-types/release-notes-templates/compiler.rst
================================================
.. _neuron-2-XX-0-compiler:
.. meta::
:description: The official release notes for the AWS Neuron SDK compiler component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.X: Neuron Compiler release notes
====================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.0 release notes home <neuron-2-XX-0-whatsnew>`
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.XX.X:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/containers.rst
================================================
.. _neuron-2-XX-0-dlc:
.. meta::
:description: The official release notes for the AWS Neuron SDK Deep Learning Containers (DLC) component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.0: Neuron Deep Learning Containers release notes
====================================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.0 release notes home <neuron-2-XX-0-whatsnew>`
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.XX.X:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/dlami.rst
================================================
.. _neuron-2-XX-0-dlami:
.. meta::
:description: The official release notes for the AWS Neuron SDK Deep Learning AWS Machine Images (DLAMIs) component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.X: Neuron Deep Learning AWS Machine Images release notes
============================================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.X release notes home <neuron-2-XX-0-whatsnew>`
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.XX.X:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/index.rst
================================================
.. _neuron-2-XX-0-whatsnew:
.. _latest-neuron-release:
.. meta::
:description: The official release notes for the AWS Neuron SDK, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.X release notes
===================================
**Date of release**: Month Day, 2026
.. toctree::
:hidden:
:maxdepth: 1
PyTorch support <nx-pytorch>
JAX support <nx-jax>
NxD Inference <nxd-inference>
NxD Training <nxd-training>
NxD Core <nxd-core>
Neuron Compiler <compiler>
NKI <nki>
Neuron Runtime <runtime>
Developer tools <tools>
Deep Learning AMIs <dlami>
Deep Learning Containers <containers>
Release artifacts <../releasecontent>
What's new?
-----------
AWS and Annapurna Labs are excited to bring you release version 2.XX.X of the Neuron SDK! In this release you'll find improvements to...
* . . .
* . . .
* . . .
.. contents:: In this release
:local:
:depth: 1
Release highlights
------------------
Version 2.XX.X brings some exciting new features! HYPE TEXT HERE
HIGHLIGHT 1
^^^^^^^^^^^
HYPE TEXT HERE
* TALKING POINT 1
* TALKING POINT 2
* . . .
USE CASE DESCRIPTION HERE
For more details, see :doc:`DOC LINK </release-notes/path/to/page>`
HIGHLIGHT 2
^^^^^^^^^^^
HYPE TEXT HERE
* TALKING POINT 1
* TALKING POINT 2
* . . .
USE CASE DESCRIPTION HERE
For more details, see :doc:`DOC LINK </release-notes/path/to/page>`
HIGHLIGHT 3
^^^^^^^^^^^
HYPE TEXT HERE
* TALKING POINT 1
* TALKING POINT 2
* . . .
USE CASE DESCRIPTION HERE
For more details, see :doc:`DOC LINK </release-notes/path/to/page>`
Other important changes
^^^^^^^^^^^^^^^^^^^^^^^
This release also includes the following improvements
* . . . LINK TO COMPONENT RELEASE NOTE PAGE
* . . . LINK TO COMPONENT RELEASE NOTE PAGE
* . . . LINK TO COMPONENT RELEASE NOTE PAGE
* . . . LINK TO COMPONENT RELEASE NOTE PAGE
Component release notes
-----------------------
Select a card below to review detailed release notes for each component of the Neuron SDK version 2.XX.X. These component release notes contain details on specific new and improved features, as well as breaking changes, bug fixes, and known issues for that component area of the Neuron SDK.
.. grid:: 1 1 2 2
:gutter: 2
.. grid-item-card::
:link: neuron-2-XX-0-pytorch
:link-type: ref
**PyTorch support** 2.XX.0 release notes
^^^
Neuron features and solutions that support the PyTorch ML framework.
+++
Supports: ``Inf2``, ``Trn1`` / ``Trn1n``, ``Trn2``
.. grid-item-card::
:link: neuron-2-XX-0-jax
:link-type: ref
**JAX support** 2.XX.0 release notes
^^^
Neuron features and solutions that support the JAX ML framework.
+++
Supports: ``Inf2``, ``Trn1`` / ``Trn1n``, ``Trn2``
.. grid-item-card::
:link: neuron-2-XX-0-nxd-training
:link-type: ref
**NxD Training** 2.XX.0 release notes
^^^
Neuron features and tools for LLM and agent ML model training.
+++
Supports: ``Trn1`` / ``Trn1n``, ``Trn2``
.. grid-item-card::
:link: neuron-2-XX-0-nxd-inference
:link-type: ref
**NxD Inference** 2.XX.0 release notes
^^^
Neuron features and tools for LLM and agent ML model inference.
+++
Supports: ``Inf2``, ``Trn1`` / ``Trn1n``, ``Trn2``
.. grid-item-card::
:link: neuron-2-XX-0-nxd-core
:link-type: ref
**NxD Core** 2.XX.0 release notes
^^^
Common features and tools for Neuron-based training and inference.
+++
Supports: ``Trn1`` / ``Trn1n``, ``Trn2``
.. grid-item-card::
:link: neuron-2-XX-0-compiler
:link-type: ref
**Neuron Compiler** 2.XX.0 release notes
^^^
The Neuron compiler for AWS Trainium and Inferentia, and its libraries and tools.
+++
Supports: ``Inf2``, ``Trn1`` / ``Trn1n``, ``Trn2``
.. grid-item-card::
:link: neuron-2-XX-0-nki
:link-type: ref
**Neuron Kernel Interface (NKI)** 2.XX.0 release notes
^^^
Neuron's Python-based programming interface for developing and optimizing Neuron kernels.
+++
Supports: ``Inf2``, ``Trn1``, ``Trn1n``
.. grid-item-card::
:link: neuron-2-XX-0-runtime
:link-type: ref
**Neuron Runtime** 2.XX.0 release notes
^^^
The Neuron kernel driver and C++ libraries for AWS Inferentia and Trainium instances.
+++
Supports: ``Inf1``, ``Inf2``, ``Trn1`` / ``Trn1n``
.. grid-item-card::
:link: neuron-2-XX-0-tools
:link-type: ref
**Neuron Developer Tools** 2.XX.0 release notes
^^^
Tools that support end-to-end development for AWS Neuron.
+++
Supports: ``Inf1``, ``Inf2``, ``Trn1`` / ``Trn1n``
.. grid-item-card::
:link: neuron-2-XX-0-dlami
:link-type: ref
**Neuron Deep Learning AWS Machine Images (DLAMIs)** 2.XX.0 release notes
^^^
AWS-specific machine images for building and deploying Neuron-based ML solutions.
+++
Supports: ``Inf1``, ``Inf2``, ``Trn1`` / ``Trn1n``
.. grid-item-card::
:link: neuron-2-XX-0-dlc
:link-type: ref
**Neuron Deep Learning Containers (DLCs)** 2.XX.0 release notes
^^^
AWS-specific container definitions for building and deploying Neuron-based ML solutions.
+++
Supports: ``Inf1``, ``Inf2``, ``Trn1`` / ``Trn1n``
.. grid-item-card::
:link: latest-neuron-release-artifacts
:link-type: ref
**Neuron 2.XX.0 release artifacts**
^^^
The libraries and packages updated in this release.
Support announcements
---------------------
This section signals the official end-of-support or end of support for specific features, tools, and APIs.
End-of-support announcements
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
*An "end-of-support (EoS)" announcement is a notification that a feature, tool, or API will not be supported in the future. Plan accordingly!*
* END-OF-SUPPORT ANNOUNCEMENT 1 (link to announcement here)
* . . .
Ending support in 2.XX.X
^^^^^^^^^^^^^^^^^^^^^^^^
"End of support" means that AWS Neuron no longer supports the feature, tool, or API indicated in the note as of this release.
* ENDING SUPPORT ANNOUNCEMENT 1 (link to announcement here)
* . . .
Previous releases
-----------------
* :doc:`Neuron 2.27.0 </release-notes/prev/2.27.0/index>`
* :doc:`Neuron 2.26.0 </release-notes/prev/2.26.0/index>`
* :doc:`Neuron 2.25.0 </release-notes/prev/2.25.0/index>`
* :doc:`Earlier releases </release-notes/prev/rn>`
* :ref:`prev-rn`
* :ref:`pre-release-content`
* :ref:`prev-n1-rn`
================================================
FILE: _content-types/release-notes-templates/nki.rst
================================================
.. _neuron-2-XX-0-nki:
.. meta::
:description: The official release notes for the AWS Neuron Kernel Interface (NKI) component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.25.0: Neuron Kernel Interace (NKI) release notes
=================================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.25.0 release notes home <neuron-2-XX-0-whatsnew>`
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.25.0:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/nx-jax.rst
================================================
.. _neuron-2-XX-0-jax:
.. meta::
:description: The official release notes for the AWS Neuron SDK JAX support component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.X: JAX support release notes
================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.25.0 release notes home <neuron-2-XX-0-whatsnew>`
Released versions
-----------------
* ``0.6.1.1.0.*``
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.25.0:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/nx-pytorch.rst
================================================
.. _neuron-2-XX-0-pytorch:
.. meta::
:description: The official release notes for AWS Neuron SDK PyTorch support, version X.XX.0. Release date: XX/XX/XXXX.
AWS Neuron SDK X.XX.0: PyTorch support release notes
====================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.0 release notes home <neuron-2-XX-0-whatsnew>`
Released versions
-----------------
* ...
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE WHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.XX.X:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/nxd-core.rst
================================================
.. _neuron-2-XX-0-nxd-core:
.. meta::
:description: The official release notes for the AWS Neuron SDK NxD Core component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.X: NxD Core release notes
=============================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.0 release notes home <neuron-2-XX-0-whatsnew>`
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.XX.X:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/nxd-inference.rst
================================================
.. _neuron-2-XX-0-nxd-inference:
.. meta::
:description: The official release notes for the AWS Neuron SDK Transformers for Inference component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.X: NxD Inference release notes
==================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.0 release notes home <neuron-2-XX-0-whatsnew>`
*
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.XX.X:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/nxd-training.rst
================================================
.. _neuron-2-XX-0-nxd-training:
.. meta::
:description: The official release notes for the AWS Neuron SDK NxD Training component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.25.0: NxD Training release notes
=================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.0 release notes home <neuron-2-XX-0-whatsnew>`
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.25.0:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/runtime.rst
================================================
.. _neuron-2-XX-0-runtime:
.. meta::
:description: The official release notes for the AWS Neuron SDK Runtime component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.X: Neuron Runtime release notes
===================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.0 release notes home <neuron-2-XX-0-whatsnew>`
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.XX.X:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _content-types/release-notes-templates/tools.rst
================================================
.. _neuron-2-XX-0-tools:
.. meta::
:description: The official release notes for the AWS Neuron SDK Developer Tools component, version X.XX.0. Release date: XX/XX/2026.
AWS Neuron SDK 2.XX.X: Developer Tools release notes
====================================================
**Date of release**: Month Day, 2026
.. contents:: In this release
:local:
:depth: 1
* Go back to the :ref:`AWS Neuron 2.XX.0 release notes home <neuron-2-XX-0-whatsnew>`
Improvements
------------
*Improvements are significant new or improved features and solutions introduced this release of the AWS Neuron SDK. Read on to learn about them!*
Feature 1
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 2
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Feature 3
^^^^^^^^^
USER-FACING DESCRIPTION OF IMPROVEMENT (WHAT WILL IT DO FOR DEV CUSTOMERS), WHY WE MADE THE IMPROVEMENT, LINK TO SUPPORTING DOC PAGE
Behavioral changes
------------------
*Behavioral changes are small, user-facing changes that you may notice after upgrading to this version.*
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HOW THE USER MAY EXPERIENCE IT, IF APPLICABLE.
* . . .
Breaking changes
----------------
*Sometimes we have to break something now to make the experience better in the longer term. Breaking changes are changes that may require you to update your own code, tools, and configurations.*
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* CHANGE DESCRIPTION SENTENCE. NOTE HWHEN THE USER MAY ENCOUNTER IT. PROVIDE A WORKAROUND, IF POSSIBLE.
* . . .
Bug fixes
---------
Here's what we fixed in 2.XX.X:
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* SHORT SENTENCE DESCRIBING BUG FIX.
* . . .
Known issues
------------
*Something doesn't work. Check here to find out if we already knew about it. We hope to fix these soon!*
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* SENTENCE DESCRIBING ISSUE AND WHEN THE USER WILL ENCOUNTER IT.
* . . .
================================================
FILE: _ext/archive.py
================================================
# This file creates a downloadable archive from each directory listed in src_dirs.
# You can modify or add additional archive_handler functions here to create additional archives.
import os, tarfile
def archive_handler(app):
old_cwd = os.getcwd()
src_dirs = ['src/examples/pytorch', 'src']
target_dirs = ['libtorch_demo', 'neuronperf']
archive_names = [name + '.tar.gz' for name in target_dirs]
for src_dir, target_dir, archive_name in zip(src_dirs, target_dirs, archive_names):
os.chdir(src_dir)
try:
os.remove(archive_name)
except OSError:
pass
with tarfile.open(archive_name, 'w:gz') as tar:
tar.add(target_dir)
os.chdir(old_cwd)
def setup(app):
app.connect('builder-inited', archive_handler)
return {
'version': '1.0',
'parallel_read_safe': True,
'parallel_write_safe': True,
}
================================================
FILE: _ext/df_tables.py
================================================
import os
from docutils.parsers.rst import Directive, directives
from docutils.parsers.rst.directives.tables import CSVTable
class DFTable(CSVTable):
CSVTable.option_spec['df-arg'] = directives.unchanged
df = None
def __init__(self, name, arguments, options, content, lineno,
content_offset, block_text, state, state_machine):
super().__init__(name, arguments, options, content, lineno,
content_offset, block_text, state, state_machine)
def get_csv_data(self):
return self.df.to_csv(index=False).splitlines(), None
def run(self):
source_file_name = self.state_machine.document.attributes["source"]
dirname = os.path.abspath(os.path.dirname(source_file_name))
os.chdir(dirname)
code = "\n".join(map(str, self.content))
ns = {}
try:
exec("\n".join( ["import numpy as np", "import pandas as pd", ]), ns)
variable_name = "df"
if self.options.get("df-var"):
variable_name = self.options.get("df-var")
exec(code, ns)
self.df = ns[variable_name]
except Exception as e:
raise self.error(str(e))
return super().run()
def setup(app):
setup.app = app
setup.config = app.config
setup.confdir = app.confdir
app.add_directive("df-table", DFTable)
metadata = {
"parallel_read_safe": True,
"parallel_write_safe": True,
"version": 0.1,
}
return metadata
================================================
FILE: _ext/local_documenter.py
================================================
import os
import sys
from sphinx.ext.autodoc import ModuleDocumenter, FunctionDocumenter
class LocalModuleDocumenter(ModuleDocumenter):
"""
Provides identical functionality to "automodule", but allows the module
function names to be overridden with the "module-name" option.
This also allows local python files to be documented as if they were
imported from an actual package by temporarily adding the directory of the
RST file to the python path.
"""
option_spec = dict(ModuleDocumenter.option_spec)
option_spec['module-name'] = lambda x = None: x
def import_object(self, *args):
"""Find modules local to the RST document directory"""
local = os.path.join(self.env.app.srcdir, os.path.dirname(self.env.docname))
sys.path.append(local)
result = super().import_object(*args)
sys.path.remove(local)
return result
def get_module_members(self):
"""Add module name override to local files"""
members = super().get_module_members()
name = self.options.module_name
if name is not None:
for member in members.values():
if callable(member.object):
setattr(member.object, 'module_name_override', name)
return members
class LocalFunctionDocumenter(FunctionDocumenter):
def format_name(self) -> str:
"""Apply module name override to local functions"""
# Use overridden module path if it is provided
if hasattr(self.object, 'module_name_override'):
self.objpath = self.object.module_name_override.split('.') + [self.objpath[-1]]
return super().format_name()
def setup(app):
app.add_autodocumenter(LocalFunctionDocumenter)
app.add_autodocumenter(LocalModuleDocumenter)
================================================
FILE: _ext/neuron_tag.py
================================================
import os
from docutils import nodes
from docutils.statemachine import ViewList
from sphinx.util.docutils import SphinxDirective
from sphinx.util.nodes import nested_parse_with_titles
# =============================================================================
# Legacy add/clear lists (used only for files NOT handled by explicit overrides)
# =============================================================================
# These lists use substring matching via in_list(). They apply ONLY when no
# explicit_override was set. As more paths get explicit overrides, entries
# here become dead code. Kept for backward compatibility with paths not yet
# explicitly overridden.
add_inf1_tag = [
'about-neuron/arch',
'archive/mxnet-neuron',
'about-neuron/announcements/index',
'archive/tensorflow/tensorflow-neuron/',
]
add_trn1_tag = [
'frameworks/neuron-customops/',
'neuron-customops/',
'frameworks/torch/inference-torch-neuronx',
'libraries/nemo-megatron/',
'libraries/nxd-training/',
]
add_trn2_tag = [
'libraries/nxd-training/',
'about-neuron/models/',
]
add_trn3_tag = [
'about-neuron/arch/neuron-hardware/neuron-core-v4',
'about-neuron/arch/neuron-hardware/trn3-arch',
]
add_neuronx_tag = [
'frameworks/torch/torch-neuronx/',
'archive/tensorflow/tensorflow-neuronx/',
'frameworks/torch/inference-torch-neuronx/',
'libraries/neuronx-distributed/',
'libraries/nxd-training',
'setup/tensorflow-neuronx',
]
clear_inf1_tag = [
'about-neuron/arch/neuron-features/neuron-caching',
'about-neuron/arch/neuron-features/eager-debug-mode',
'about-neuron/arch/neuron-features/collective-communication-operations',
'about-neuron/arch/neuron-features/dynamic-shapes',
'about-neuron/arch/neuron-features/control-flow',
'about-neuron/arch/neuron-features/custom-c++-operators',
'about-neuron/arch/neuron-features/collective-communication',
'about-neuron/arch/neuron-features/rounding-modes',
'about-neuron/arch/neuron-hardware/trn1-arch',
'about-neuron/arch/neuron-hardware/inf2-arch',
'about-neuron/arch/neuron-hardware/inferentia2',
'about-neuron/arch/neuron-hardware/trainium',
'about-neuron/arch/neuron-hardware/neuron-core-v2',
'about-neuron/arch/neuron-hardware/trn2-arch',
'about-neuron/arch/neuron-hardware/trn3-arch',
'about-neuron/arch/neuron-hardware/neuron-core-v3',
'about-neuron/arch/neuron-hardware/neuron-core-v4',
'about-neuron/benchmarks/trn1-performance',
'about-neuron/benchmarks/trn1/',
'about-neuron/benchmarks/inf2/inf2-performance',
'about-neuron/faq/training/',
'about-neuron/models/inference-inf2-trn1-samples',
'about-neuron/models/training-trn1-samples',
'about-neuron/models/training-inference-trn2-samples',
'about-neuron/appnotes/neuronx-cc/neuronx-cc-training-mixed-precision',
'about-neuron/appnotes/transformers-neuronx/generative-llm-inference-with-neuron',
'about-neuron/appnotes/torch-neuronx/torch-neuronx-dataparallel-app-note',
'about-neuron/calculator/neuron-calculator',
'about-neuron/announcements/neuron2.x/dlami-pytorch-introduce',
'about-neuron/announcements/neuron2.x/sm-training-trn1-introduce',
'about-neuron/announcements/neuron2.x/sm-training-dlc-2.9.1',
'devflows/training',
'devflows/inference/byoc-hosting-devflow-inf2',
'compiler/neuronx-cc/',
'about-neuron/appnotes/perf/neuronx-cc/',
'frameworks/torch/torch-neuronx/',
'frameworks/torch/training',
'frameworks/torch/inference-torch-neuronx',
'archive/tensorflow/tensorflow-neuronx/',
'archive/tensorflow/tensorflow-neuronx-inference',
'frameworks/torch/torch-neuronx/transformers-neuronx/readme',
'release-notes/neuron-cc/index',
'release-notes/runtime/aws-neuronx-collectives/',
'release-notes/torch/torch-neuronx/',
'release-notes/torch/transformers-neuronx/index',
'release-notes/tensorflow/tensorflow-neuronx/',
'release-notes/compiler/neuronx-cc/',
'tools/tutorials/tutorial-tensorboard-scalars-mnist',
'tools/tutorials/tutorial-neuron-monitor-mnist',
'tools/tensorboard/getting-started-tensorboard-neuronx-plugin',
'tools/neuron-sys-tools/nccom-test',
'setup/torch-neuronx',
'setup/tensorflow-neuronx',
'setup/neuron-setup/tensorflow/neuronx/',
'setup/neuron-setup/pytorch/neuronx/',
'nki/',
'frameworks/jax/',
'libraries/nxd-training/',
'/release-notes/components/nki',
'/release-notes/components/nki-lib',
'/release-notes/components/compiler'
]
clear_inf2_tag = [
'frameworks/torch/torch-neuronx/training',
'frameworks/torch/training',
'archive/torch-neuron/inference-torch-neuron',
'archive/tensorflow/tensorflow-neuron-inference',
'frameworks/jax/',
'about-neuron/arch/neuron-hardware/trn1-arch',
'about-neuron/arch/neuron-hardware/trainium',
'about-neuron/arch/neuron-hardware/trn2-arch',
'about-neuron/arch/neuron-hardware/trn3-arch',
'about-neuron/arch/neuron-hardware/neuron-core-v3',
'about-neuron/arch/neuron-hardware/neuron-core-v4',
'about-neuron/arch/neuron-features/logical-neuroncore-config',
'about-neuron/benchmarks/trn1/trn1-inference-performance',
'about-neuron/benchmarks/trn1/trn1-training-performance',
'about-neuron/models/training-trn1-samples',
'about-neuron/models/training-inference-trn2-samples',
'about-neuron/announcements/neuron2.x/announce-neuron-trn2',
'neuronx-distributed/nxd-training',
'libraries/nxd-training/',
'tools/neuron-sys-tools/nccom-test',
'release-notes/runtime/aws-neuronx-collectives/',
]
clear_trn1_tag = [
'about-neuron/arch/neuron-hardware/inf2-arch',
'about-neuron/arch/neuron-hardware/inferentia2',
'about-neuron/arch/neuron-hardware/trn2-arch',
'about-neuron/arch/neuron-hardware/trn3-arch',
'about-neuron/arch/neuron-hardware/trainium2',
'about-neuron/arch/neuron-hardware/neuron-core-v3',
'about-neuron/arch/neuron-hardware/neuron-core-v4',
'about-neuron/benchmarks/inf2/inf2-performance',
'about-neuron/models/training-inference-trn2-samples',
]
clear_trn2_tag = [
'archive/tensorflow/',
'libraries/transformers-neuronx/',
'about-neuron/arch/neuron-hardware/trn1-arch',
'about-neuron/arch/neuron-hardware/trainium',
'about-neuron/arch/neuron-hardware/neuron-core-v2',
'about-neuron/arch/neuron-hardware/neuron-core-v4',
'about-neuron/arch/neuron-hardware/trn3-arch',
'about-neuron/benchmarks/',
'about-neuron/benchmarks/trn1/',
'about-neuron/benchmarks/inf2/inf2-performance',
'about-neuron/models/inference-inf2-trn1-samples',
'about-neuron/models/training-trn1-samples',
'neuron-customops/programming-guide/custom-c++-operators-devguide'
]
clear_trn3_tag = [
'archive/tensorflow/',
'libraries/transformers-neuronx/',
'about-neuron/arch/neuron-hardware/trn1-arch',
'about-neuron/arch/neuron-hardware/trainium',
'about-neuron/arch/neuron-hardware/neuron-core-v2',
'about-neuron/arch/neuron-hardware/neuron-core-v3',
'about-neuron/benchmarks/',
'about-neuron/benchmarks/trn1/',
'about-neuron/benchmarks/inf2/inf2-performance',
'about-neuron/models/inference-inf2-trn1-samples',
'about-neuron/models/training-trn1-samples',
'libraries/neuronx-distributed/context_parallelism_overview',
'about-neuron/appnotes/',
'neuron-customops/programming-guide/custom-c++-operators-devguide'
]
# Neuron 1.x / NeuronCore v1 era content — clear all non-Inf1 tags
clear_nc_v2_tag = [
'tools/tutorials/tutorial-neuron-check-model',
'tools/tutorials/tutorial-neuron-gatherinfo',
'tools/tutorials/getting-started-tensorboard-neuron-plugin',
'tools/tensorboard/getting-started-tensorboard-neuron-plugin',
'tools/helper-tools/tutorial-neuron-check-model',
'tools/helper-tools/tutorial-neuron-gatherinfo',
'about-neuron/appnotes/neuron-cc/mixed-precision',
'about-neuron/appnotes/perf/neuron-cc/',
'about-neuron/appnotes/neuron1x/',
'about-neuron/appnotes/torch-neuron/',
'about-neuron/arch/neuron-hardware/inf1-arch',
'about-neuron/arch/neuron-hardware/inferentia',
'about-neuron/arch/neuron-hardware/neuron-core-v1',
'about-neuron/arch/neuron-features/neuroncore-pipeline',
'about-neuron/announcements/neuron1.x/',
'about-neuron/quick-start/mxnet-neuron',
'about-neuron/benchmarks/inf1/',
'about-neuron/faq/inference/',
'about-neuron/models/inference-inf1-samples',
'containers/dlc-then-ec2-devflow',
'containers/dlc-then-ecs-devflow',
'containers/dlc-then-eks-devflow',
'containers/container-sm-hosting-devflow',
'containers/rn',
'containers/tutorials/k8s-neuron-scheduler',
'compiler/neuron-cc/',
'release-notes/mxnet-neuron/',
'release-notes/torch/torch-neuron/',
'release-notes/tensorflow/tensorflow-neuron/',
'release-notes/compiler/neuron-cc/',
'release-notes/neuron1/',
'archive/torch-neuron/',
'archive/torch-neuron/inference-torch-neuron',
'archive/tensorflow/tensorflow-neuron/',
'archive/tensorflow/tensorflow-neuron-inference',
'archive/mxnet-neuron/',
'setup/tensorflow-neuron',
'setup/torch-neuron',
'setup/mxnet-neuron',
'setup/neuron-setup/pytorch/neuron/',
'setup/neuron-setup/mxnet/neuron/ubuntu/',
'setup/neuron-setup/mxnet/neuron/amazon-linux/',
'setup/neuron-setup/tensorflow/neuron/ubuntu/',
'setup/neuron-setup/tensorflow/neuron/amazon-linux/',
]
# Top-level directories used for initial tag assignment
NEURON1_DIRS = ['n1']
COMMON_DIRS = [
'tools', 'neuron-runtime', 'release-notes', 'containers', 'compiler',
'frameworks', 'src', 'about-neuron', 'setup', 'devflows', 'dlami', 'libraries',
]
TEXT_TEMPLATE = '**This document is relevant for**: '
# =============================================================================
# Hardware architecture page map (exact docname → instance list)
# =============================================================================
HW_ARCH_MAP = {
'about-neuron/arch/neuron-hardware/inf1-arch': ['Inf1'],
'about-neuron/arch/neuron-hardware/inf2-arch': ['Inf2'],
'about-neuron/arch/neuron-hardware/inferentia': ['Inf1'],
'about-neuron/arch/neuron-hardware/inferentia2': ['Inf2'],
'about-neuron/arch/neuron-hardware/neuron-core-v1': ['Inf1'],
'about-neuron/arch/neuron-hardware/neuron-core-v2': ['Inf2', 'Trn1'],
'about-neuron/arch/neuron-hardware/neuron-core-v3': ['Trn2'],
'about-neuron/arch/neuron-hardware/neuron-core-v4': ['Trn3'],
'about-neuron/arch/neuron-hardware/trainium': ['Trn1'],
'about-neuron/arch/neuron-hardware/trainium2': ['Trn2'],
'about-neuron/arch/neuron-hardware/trainium3': ['Trn3'],
'about-neuron/arch/neuron-hardware/trn1-arch': ['Trn1'],
'about-neuron/arch/neuron-hardware/trn2-arch': ['Trn2'],
'about-neuron/arch/neuron-hardware/trn3-arch': ['Trn3'],
}
# NxD Core training-specific pages (no Inf2)
NXD_CORE_TRAINING_PAGES = [
'libraries/neuronx-distributed/index-training',
'libraries/neuronx-distributed/developer-guide-training',
'libraries/neuronx-distributed/api-reference-guide-training',
'libraries/neuronx-distributed/tp_developer_guide',
'libraries/neuronx-distributed/pp_developer_guide',
'libraries/neuronx-distributed/ptl_developer_guide',
'libraries/neuronx-distributed/save_load_developer_guide',
'libraries/neuronx-distributed/activation_memory_reduction',
'libraries/neuronx-distributed/activation_memory_reduction_developer_guide',
'libraries/neuronx-distributed/standard_mixed_precision',
'libraries/neuronx-distributed/tensor_parallelism_overview',
'libraries/neuronx-distributed/pipeline_parallelism_overview',
'libraries/neuronx-distributed/lora_finetune_developer_guide',
'libraries/neuronx-distributed/model_optimizer_wrapper_developer_guide',
'libraries/neuronx-distributed/context_parallelism_overview',
]
def _in_list(cur_file, file_list):
"""Return True if any entry in file_list is a substring of cur_file."""
return any(entry in cur_file for entry in file_list)
def _splitall(path):
"""Split a path into all its components."""
parts = []
while True:
head, tail = os.path.split(path)
if head == path:
parts.insert(0, head)
break
elif tail == path:
parts.insert(0, tail)
break
else:
path = head
parts.insert(0, tail)
return parts, len(parts)
def _get_explicit_override(cur_file):
"""Return (instances, True) if cur_file has an explicit CSV-based override,
or (None, False) otherwise.
Rules are evaluated top-to-bottom. More specific paths must come AFTER
broader paths so they can override them (last match wins).
"""
# --- Libraries -----------------------------------------------------------
# NxD Core = Inf2, Trn1, Trn2 (default for all neuronx-distributed pages)
if cur_file.startswith('libraries/neuronx-distributed/'):
result = ['Inf2', 'Trn1', 'Trn2']
# Training-specific pages drop Inf2
if cur_file in NXD_CORE_TRAINING_PAGES:
result = ['Trn1', 'Trn2']
if cur_file.startswith('libraries/neuronx-distributed/tutorials/training') or \
cur_file.startswith('libraries/neuronx-distributed/tutorials/finetune'):
result = ['Trn1', 'Trn2']
return result, True
if cur_file.startswith('libraries/transformers-neuronx/'):
return ['Inf2', 'Trn1'], True
if cur_file.startswith('libraries/nxd-training/'):
return ['Trn1', 'Trn2'], True
# vLLM must come before general nxd-inference
if cur_file.startswith('libraries/nxd-inference/vllm/'):
return ['Trn2', 'Trn3'], True
if cur_file.startswith('libraries/nxd-inference/'):
return ['Inf2', 'Trn1', 'Trn2'], True
if cur_file.startswith('libraries/nemo-megatron/'):
return ['Trn1', 'Trn2'], True
# --- NKI -----------------------------------------------------------------
if cur_file.startswith('nki/'):
return ['Trn2', 'Trn3'], True
# --- CustomOps -----------------------------------------------------------
if cur_file.startswith('neuron-customops/'):
return ['Inf2', 'Trn1'], True
# --- Frameworks ----------------------------------------------------------
if cur_file.startswith('frameworks/jax/'):
return ['Trn2', 'Trn3'], True
# TensorFlow NeuronX (must come before TensorFlow Neuron check)
if 'tensorflow/tensorflow-neuronx' in cur_file:
return ['Inf2', 'Trn1'], True
# TensorFlow Neuron (Inf1)
if 'tensorflow/tensorflow-neuron' in cur_file and 'neuronx' not in cur_file:
return ['Inf1'], True
# TorchNeuron native PyTorch (must come before torch-neuronx check)
if 'torch/pytorch-native' in cur_file:
return ['Trn2', 'Trn3'], True
# PyTorch NeuronX (Torch/XLA)
if 'torch/torch-neuronx' in cur_file:
return ['Inf2', 'Trn1', 'Trn2'], True
# PyTorch NeuronX top-level pages (not in torch-neuronx/ subdir)
if cur_file in ['frameworks/torch/inference-torch-neuronx',
'frameworks/torch/training-torch-neuronx',
'frameworks/torch/training',
'frameworks/torch/inference']:
return ['Inf2', 'Trn1', 'Trn2'], True
# PyTorch Neuron (Inf1)
if 'torch/torch-neuron' in cur_file and 'neuronx' not in cur_file:
return ['Inf1'], True
if cur_file == 'archive/torch-neuron/inference-torch-neuron':
return ['Inf1'], True
# MXNet
if 'mxnet-neuron' in cur_file:
return ['Inf1'], True
# --- Neuron Runtime ------------------------------------------------------
# Collectives (more specific, must come after general runtime)
if cur_file.startswith('neuron-runtime/about/collectives') or \
cur_file in ['neuron-runtime/explore/internode-collective-comm',
'neuron-runtime/explore/intranode-collective-comm',
'neuron-runtime/explore/compute-comm-overlap']:
return ['Trn1', 'Trn2', 'Trn3'], True
if cur_file.startswith('neuron-runtime/'):
return ['Inf2', 'Trn1', 'Trn2', 'Trn3'], True
# --- Compiler ------------------------------------------------------------
if cur_file.startswith('compiler/error-codes/'):
return ['Inf2', 'Trn1', 'Trn2', 'Trn3'], True
if cur_file == 'compiler/neuron-cc' or cur_file.startswith('compiler/neuron-cc/'):
return ['Inf1'], True
if cur_file == 'compiler/neuronx-cc' or cur_file.startswith('compiler/neuronx-cc/'):
return ['Inf2', 'Trn1', 'Trn2', 'Trn3'], True
if cur_file == 'neuron-customops/programming-guide' or cur_file.startswith('neuron-customops/programming-guide'):
return ['Inf2', 'Trn1'], True
# --- Setup ---------------------------------------------------------------
if cur_file.startswith('setup/install-templates/inf1/'):
return ['Inf1'], True
if cur_file.startswith('setup/install-templates/inf2/'):
return ['Inf2'], True
if cur_file.startswith('setup/install-templates/trn1/') or \
cur_file == 'setup/install-templates/launch-trn1-dlami':
return ['Trn1'], True
if cur_file in ['setup/setup-neuron', 'setup/torch-neuron', 'setup/torch-neuron-ubuntu20']:
return ['Inf1'], True
if cur_file.startswith('setup/neuron-setup/pytorch/neuronx/'):
return ['Inf2', 'Trn1', 'Trn2'], True
if cur_file.startswith('setup/neuron-setup/tensorflow/neuronx/'):
return ['Inf2', 'Trn1'], True
if cur_file.startswith('setup/neuron-setup/pytorch/neuron/'):
return ['Inf1'], True
if cur_file.startswith('setup/neuron-setup/tensorflow/neuron/'):
return ['Inf1'], True
if cur_file == 'setup/jax-neuronx':
return ['Trn2', 'Trn3'], True
if cur_file == 'setup/torch-neuronx':
return ['Inf2', 'Trn1', 'Trn2'], True
if cur_file == 'setup/tensorflow-neuronx':
return ['Inf2', 'Trn1'], True
if cur_file == 'setup/tensorflow-neuron':
return ['Inf1'], True
return None, False
def _get_page_override(cur_file):
"""Return (instances, True) for page-specific overrides that don't fit
neatly into _get_explicit_override (devflows, containers, tools, about-neuron, etc.).
"""
# --- Devflows ------------------------------------------------------------
if cur_file == 'devflows/inference/byoc-hosting-devflow-inf2':
return ['Inf2'], True
if cur_file == 'devflows/inference/ec2-then-ec2-devflow-inf2':
return ['Inf2'], True
if cur_file == 'devflows/parallelcluster-flows':
return ['Trn1', 'Trn2'], True
if cur_file.startswith('devflows/training/batch/') or \
cur_file.startswith('devflows/training/ec2/') or \
cur_file.startswith('devflows/training/parallelcluster/') or \
cur_file.startswith('devflows/training/sm-devflow/'):
return ['Trn1', 'Trn2', 'Trn3'], True
if cur_file.startswith('devflows/plugins/npd'):
return ['Inf2', 'Trn1', 'Trn2'], True
# --- Containers ----------------------------------------------------------
# OCI Hooks
if 'tutorial-oci-hook' in cur_file:
return ['Inf1', 'Inf2', 'Trn1', 'Trn2'], True
# DRA
if cur_file == 'containers/neuron-dra' or cur_file.startswith('containers/files/'):
return ['Trn2', 'Trn3'], True
if cur_file == 'containers/how-to/how-to-ultraserver':
return ['Trn2', 'Trn3'], True
# DLC quickstarts
if cur_file == 'containers/get-started/quickstart-configure-deploy-dlc':
return ['Trn2', 'Trn3'], True
if cur_file == 'containers/get-started/quickstart-pytorch-inference-dlc':
return ['Inf2', 'Trn1', 'Trn2', 'Trn3'], True
# Inf1-era container content
if cur_file == 'containers/tutorial-docker-runtime1.0':
return ['Inf1'], True
if cur_file == 'containers/container-deployment-flows' or \
cur_file.startswith('containers/docker-example/inference/') or \
cur_file.startswith('containers/docker-example/v1/') or \
cur_file == 'containers/ec2-then-ec2-devflow' or \
cur_file == 'containers/neo-then-hosting-devflow':
return ['Inf1'], True
# Container training/inference tutorials and docker examples
if cur_file.startswith('containers/docker-example/training/'):
return ['Trn1', 'Trn2', 'Trn3'], True
if cur_file.startswith('containers/tutorials/inference/'):
return ['Inf1'], True
if cur_file.startswith('containers/tutorials/training/'):
return ['Trn1', 'Trn2', 'Trn3'], True
# Neuron Monitor Container
if cur_file == 'containers/tutorials/k8s-neuron-monitor':
return ['Inf2', 'Trn1', 'Trn2'], True
# Node Problem Detector
if cur_file.startswith('containers/tutorials/k8s-neuron-problem-detector'):
return ['Inf2', 'Trn1', 'Trn2'], True
# --- Tools ---------------------------------------------------------------
# TensorBoard plugin (End Of Support)
if cur_file.startswith('tools/tensorboard/getting-started-tensorboard-neuronx') or \
cur_file == 'tools/tutorials/tutorial-tensorboard-scalars-mnist' or \
cur_file == 'tools/tutorials/torch-neuronx-profiling-with-tb':
return ['Inf2', 'Trn1'], True
# --- Announcements -------------------------------------------------------
if cur_file.startswith('about-neuron/announcements/'):
return [], True
# --- Hardware architecture -----------------------------------------------
if cur_file in HW_ARCH_MAP:
return HW_ARCH_MAP[cur_file], True
# --- Arch features -------------------------------------------------------
if cur_file == 'about-neuron/arch/neuron-features/custom-c++-operators':
return ['Inf2', 'Trn1'], True
if cur_file == 'about-neuron/arch/neuron-features/logical-neuroncore-config':
return ['Trn2', 'Trn3'], True
# --- Appnotes ------------------------------------------------------------
if cur_file == 'about-neuron/appnotes/neuronx-distributed/introducing-nxd-inference':
return ['Inf2', 'Trn1', 'Trn2'], True
if cur_file == 'about-neuron/appnotes/neuronx-distributed/introducing-nxdt-training':
return ['Trn1', 'Trn2'], True
if cur_file.startswith('about-neuron/appnotes/torch-neuronx/'):
return ['Inf2', 'Trn1', 'Trn2'], True
if cur_file.startswith('about-neuron/appnotes/transformers-neuronx/'):
return ['Inf2', 'Trn1'], True
if cur_file == 'about-neuron/appnotes/neuronx-cc/neuronx-cc-training-mixed-precision':
return ['Trn1', 'Trn2', 'Trn3'], True
if cur_file.startswith('about-neuron/appnotes/neuron1x/'):
return ['Inf1'], True
# --- Benchmarks ----------------------------------------------------------
if cur_file == 'about-neuron/benchmarks/index':
return ['Inf1', 'Inf2', 'Trn1', 'Trn2', 'Trn3'], True
# --- Quick-start ---------------------------------------------------------
if cur_file == 'about-neuron/quick-start/tensorflow-neuron':
return ['Inf1'], True
if cur_file in ['about-neuron/quick-start/torch-neuron',
'about-neuron/quick-start/torch-neuron-tab-training']:
return ['Inf1'], True
if cur_file.startswith('about-neuron/quick-start/tab-inference-torch-neuronx'):
return ['Inf2', 'Trn1', 'Trn2'], True
if cur_file.startswith('about-neuron/quick-start/tab-inference-torch-neuron') and 'neuronx' not in cur_file:
return ['Inf1'], True
if cur_file.startswith('about-neuron/quick-start/tab-inference-tensorflow-neuronx'):
return ['Inf2', 'Trn1'], True
if cur_file.startswith('about-neuron/quick-start/tab-inference-tensorflow-neuron') and 'neuronx' not in cur_file:
return ['Inf1'], True
return None, False
class NeuronTag(SphinxDirective):
def run(self):
cur_file = self.env.docname
path_split, path_len = _splitall(cur_file)
# Landing page gets no tag
if path_split[0] == 'index':
return self._render('')
# Step 1: Assign default instances based on top-level directory
return_instances = []
if path_split[0] in NEURON1_DIRS:
return_instances = ['Inf1']
elif path_split[0] in COMMON_DIRS:
return_instances = ['Inf1', 'Inf2', 'Trn1', 'Trn2', 'Trn3']
# Step 2: Check explicit overrides (CSV-based, highest priority)
explicit_override = False
result, matched = _get_explicit_override(cur_file)
if matched:
return_instances = result
explicit_override = True
if not explicit_override:
result, matched = _get_page_override(cur_file)
if matched:
return_instances = result
explicit_override = True
# Step 3: Directory-based inference/training heuristic
if not explicit_override:
if path_len >= 2:
parent_dir = path_split[path_len - 2]
if parent_dir == 'inference':
return_instances = ['Inf1']
elif parent_dir == 'training':
return_instances = ['Trn1', 'Trn2', 'Trn3']
# Step 4: Legacy add/clear tag lists (only for non-overridden files)
if not explicit_override:
if _in_list(cur_file, add_trn1_tag):
if 'Trn1' not in return_instances:
return_instances.extend(['Trn1', 'Trn2', 'Trn3', 'Inf2'])
if _in_list(cur_file, add_trn2_tag):
if 'Trn2' not in return_instances:
return_instances.extend(['Trn2', 'Trn3'])
if _in_list(cur_file, add_trn3_tag):
if 'Trn3' not in return_instances:
return_instances.append('Trn3')
if _in_list(cur_file, add_neuronx_tag):
if 'Trn1' not in return_instances:
return_instances.extend(['Trn1', 'Trn2', 'Trn3', 'Inf2'])
if _in_list(cur_file, add_inf1_tag):
if 'Inf1' not in return_instances:
return_instances.append('Inf1')
if _in_list(cur_file, clear_nc_v2_tag):
for tag in ['Trn1', 'Trn2', 'Trn3', 'Inf2']:
if tag in return_instances:
return_instances.remove(tag)
if _in_list(cur_file, clear_trn1_tag):
if 'Trn1' in return_instances:
return_instances.remove('Trn1')
if _in_list(cur_file, clear_trn2_tag):
if 'Trn2' in return_instances:
return_instances.remove('Trn2')
if _in_list(cur_file, clear_trn3_tag):
if 'Trn3' in return_instances:
return_instances.remove('Trn3')
if _in_list(cur_file, clear_inf1_tag):
if 'Inf1' in return_instances:
return_instances.remove('Inf1')
if _in_list(cur_file, clear_inf2_tag):
if 'Inf2' in return_instances:
return_instances.remove('Inf2')
# Step 5: Generate output
return_instances = sorted(set(return_instances))
if return_instances:
text = TEXT_TEMPLATE + ', '.join('``' + i + '``' for i in return_instances)
else:
text = ''
return self._render(text)
def _render(self, text):
"""Parse RST text and return docutils nodes."""
rst = ViewList()
rst.append(text, "neuron-tag", 1)
node = nodes.section()
node.document = self.state.document
nested_parse_with_titles(self.state, rst, node)
return node.children
def setup(app):
app.add_directive("neuron-tag", NeuronTag)
return {
'version': '0.2',
'parallel_read_safe': True,
'parallel_write_safe': True,
}
================================================
FILE: _ext/release-notes-automation-spec.md
================================================
# Release Notes Review Automation Specification
## Overview
This specification defines a GitHub Action that automatically reviews release notes files in pull requests using Amazon Q CLI to ensure they meet quality standards defined in the release notes writing guidelines.
## Purpose
Automate the review of release notes changes to:
- Ensure consistency and quality across all release notes
- Catch common issues before human review
- Provide immediate feedback to PR authors
- Reduce manual review burden on documentation team
## Scope
### In Scope
- PRs labeled with "release-notes"
- RST files under `/release-notes/components/` directory
- Files that have been modified in the PR (not just added to context)
- Automated review using Q CLI with release notes guidelines
- Posting review feedback as PR comments
### Out of Scope
- Release notes files outside `/release-notes/components/`
- Non-RST files
- PRs without the "release-notes" label
- Manual approval/rejection of PRs (action only provides feedback)
## Requirements
### Functional Requirements
#### FR1: PR Detection and Filtering
- **FR1.1**: Action triggers on pull request events (opened, synchronize, labeled)
- **FR1.2**: Action only runs when PR has "release-notes" label
- **FR1.3**: Action identifies all changed RST files in `/release-notes/components/` directory
#### FR2: File Analysis
- **FR2.1**: Action reads content of each changed RST file
- **FR2.2**: Action loads release notes guidelines from `_ext/release-notes-context.md`
- **FR2.3**: Action processes files individually to provide file-specific feedback
#### FR3: Q CLI Integration
- **FR3.1**: Action invokes Amazon Q CLI with appropriate context
- **FR3.2**: Action provides Q CLI with:
- Release notes guidelines from `_ext/release-notes-context.md`
- Content of the changed RST file
- Instruction to review against guidelines
- **FR3.3**: Action captures Q CLI output for each file
#### FR4: Review Feedback
- **FR4.1**: Action formats Q CLI feedback into readable PR comment
- **FR4.2**: Action posts comment to PR with review results
- **FR4.3**: Comment includes:
- List of files reviewed
- Issues found per file (using format from guidelines)
- Suggested improvements
- Link to full guidelines document
- **FR4.4**: If no issues found, action posts positive confirmation
#### FR5: Error Handling
- **FR5.1**: Action handles Q CLI failures gracefully
- **FR5.2**: Action reports when no RST files are found in scope
- **FR5.3**: Action logs errors for debugging without failing the PR
### Non-Functional Requirements
#### NFR1: Performance
- Action completes review within 5 minutes for typical PRs (1-5 files)
- Action processes files in parallel when possible
#### NFR2: Security
- Action uses GitHub secrets for Q CLI credentials
- Action has read-only access to repository
- Action has write access only to PR comments
#### NFR3: Maintainability
- Action configuration is version controlled in `.github/workflows/`
- Action uses official Q CLI container/action when available
- Action logic is simple and well-documented
## User Stories
### US1: Automatic Review Trigger
**As a** documentation contributor
**I want** the review action to run automatically when I label my PR
**So that** I get immediate feedback without manual intervention
**Acceptance Criteria:**
- Action triggers when "release-notes" label is added
- Action runs on subsequent commits to labeled PR
- Action does not run on PRs without the label
### US2: Targeted File Review
**As a** documentation contributor
**I want** only my changed release notes files to be reviewed
**So that** I get relevant feedback without noise from unchanged files
**Acceptance Criteria:**
- Only files in `/release-notes/components/*.rst` are reviewed
- Only files modified in the PR are analyzed
- Files in other directories are ignored
### US3: Clear Feedback
**As a** documentation contributor
**I want** clear, actionable feedback on my release notes
**So that** I know exactly what to improve
**Acceptance Criteria:**
- Feedback follows the format specified in guidelines
- Each issue includes: original text, problem, example rewrite, action items
- Feedback is posted as a PR comment
- Comment includes link to full guidelines
### US4: No False Failures
**As a** documentation contributor
**I want** the action to provide feedback without blocking my PR
**So that** I can address issues without being blocked by automation
**Acceptance Criteria:**
- Action never fails the PR check
- Action always succeeds even if issues are found
- Issues are reported as comments, not check failures
## Technical Design
### GitHub Action Workflow
**File Location:** `.github/workflows/release-notes-review.yml`
**Trigger Events:**
```yaml
on:
pull_request:
types: [opened, synchronize, labeled]
paths:
- 'release-notes/components/**/*.rst'
```
**Workflow Steps:**
1. **Check Label**
- Verify PR has "release-notes" label
- Exit gracefully if label not present
2. **Get Changed Files**
- Use GitHub API to get list of changed files
- Filter for `release-notes/components/**/*.rst`
- Exit if no matching files found
3. **Setup Q CLI**
- Install/configure Amazon Q CLI
- Authenticate using GitHub secrets
4. **Load Guidelines**
- Read `_ext/release-notes-context.md`
- Prepare as context for Q CLI
5. **Review Each File**
- For each changed RST file:
- Read file content
- Invoke Q CLI with prompt:
```
Review the following release notes file against the guidelines provided.
Guidelines: [content from release-notes-context.md]
File: [filename]
Content: [file content]
Provide feedback using the review format specified in the guidelines.
Focus on: customer visibility, documentation links, impact clarity,
specific conditions, and actionable information.
```
- Capture Q CLI response
6. **Format Feedback**
- Combine all file reviews into single comment
- Format as markdown with sections per file
- Include summary at top
7. **Post Comment**
- Post formatted feedback as PR comment
- Include link to guidelines
- Tag PR author
### Q CLI Prompt Template
```markdown
You are reviewing release notes for the AWS Neuron SDK. Review the following
file against the release notes writing guidelines.
GUIDELINES:
[Full content of _ext/release-notes-context.md]
FILE TO REVIEW: {filename}
CONTENT:
{file_content}
INSTRUCTIONS:
1. Review the content against all guidelines
2. Identify issues using the review format from the guidelines
3. For each issue, provide:
- Issue number and title
- Original text
- Problem description
- Phrasing problem (if applicable)
- Example rewrite
- Specific action items
4. If no issues found, state "No issues found - release notes meet guidelines"
Focus especially on:
- Customer-visible language (no internal code names)
- Documentation URLs for all new features
- Specific conditions (not vague language)
- Clear impact statements
- Proper categorization (breaking changes vs bug fixes)
- Migration guidance for breaking changes
gitextract_u554eb9v/
├── .github/
│ ├── ISSUE_TEMPLATE/
│ │ ├── bug-report.yml
│ │ ├── config.yml
│ │ ├── documentation.yml
│ │ └── feature-request.yml
│ ├── pull_request_template.md
│ ├── stale_issue_mark_close_workflow.yml
│ └── workflows/
│ ├── acknowledge-new-issue.yml
│ └── auto-label-issues.yml
├── .gitignore
├── .readthedocs.yml
├── CODEOWNERS
├── CONTRIBUTING.md
├── Dockerfile
├── LICENSE-DOCUMENTATION
├── LICENSE-SAMPLECODE
├── LICENSE-SUMMARY-DOCS-SAMPLES
├── Makefile
├── README.md
├── _backup-setup/
│ └── neuron-setup/
│ ├── multiframework/
│ │ ├── multi-framework-ubuntu22-neuron-dlami.rst
│ │ └── multi-framework-ubuntu24-neuron-dlami.rst
│ └── pytorch/
│ ├── neuron/
│ │ ├── amazon-linux/
│ │ │ ├── torch-neuron-al2-base-dlami.rst
│ │ │ ├── torch-neuron-al2-pytorch-dlami.rst
│ │ │ ├── torch-neuron-al2.rst
│ │ │ └── torch-neuron-al2023.rst
│ │ └── ubuntu/
│ │ ├── torch-neuron-ubuntu20-base-dlami.rst
│ │ ├── torch-neuron-ubuntu20-pytorch-dlami.rst
│ │ ├── torch-neuron-ubuntu20.rst
│ │ └── torch-neuron-ubuntu22.rst
│ └── neuronx/
│ ├── amazon-linux/
│ │ ├── torch-neuronx-al2-base-dlami.rst
│ │ ├── torch-neuronx-al2-pytorch-dlami.rst
│ │ ├── torch-neuronx-al2.rst
│ │ └── torch-neuronx-al2023.rst
│ └── ubuntu/
│ ├── torch-neuronx-ubuntu20-base-dlami.rst
│ ├── torch-neuronx-ubuntu20-pytorch-dlami.rst
│ ├── torch-neuronx-ubuntu20.rst
│ ├── torch-neuronx-ubuntu22.rst
│ └── torch-neuronx-ubuntu24.rst
├── _content-types/
│ ├── conceptual-deep-dive.rst
│ ├── model-card.rst
│ ├── procedural-how-to.rst
│ ├── procedural-tutorial.ipynb
│ ├── reference-kernel-api.rst
│ └── release-notes-templates/
│ ├── compiler.rst
│ ├── containers.rst
│ ├── dlami.rst
│ ├── index.rst
│ ├── nki.rst
│ ├── nx-jax.rst
│ ├── nx-pytorch.rst
│ ├── nxd-core.rst
│ ├── nxd-inference.rst
│ ├── nxd-training.rst
│ ├── runtime.rst
│ └── tools.rst
├── _ext/
│ ├── archive.py
│ ├── df_tables.py
│ ├── local_documenter.py
│ ├── neuron_tag.py
│ ├── release-notes-automation-spec.md
│ ├── release-notes-context.md
│ ├── sphinx_plotly_directive.py
│ └── symlink.py
├── _static/
│ └── css/
│ ├── custom.css
│ └── custom.css.new
├── _templates/
│ ├── recentposts.html
│ ├── search-field.html
│ ├── search-google.html
│ └── search.html
├── _utilities/
│ ├── JIRA_SETUP_QUICKSTART.md
│ ├── add_meta.py
│ ├── audit_frameworks.py
│ ├── check_urls.sh
│ ├── create_sitemap.py
│ ├── format_build_logs.py
│ ├── inject_archive_meta.py
│ ├── metadata_schema.yaml
│ ├── migrate_setup_content.py
│ ├── old-nki-apis.txt
│ └── setup_jira_token.sh
├── about-neuron/
│ ├── amazonq-getstarted.rst
│ ├── announcements/
│ │ ├── index.rst
│ │ ├── neuron1.x/
│ │ │ ├── announce-eol-mx-before-1-5.rst
│ │ │ ├── announce-eol-pt-1-5.rst
│ │ │ ├── announce-eol-pt-before-1-8.rst
│ │ │ ├── announce-eol-tf-before-2-5.rst
│ │ │ ├── announce-eol-tf-before-2-7.rst
│ │ │ ├── announcements.rst
│ │ │ ├── eol-ncgs-env_2.rst
│ │ │ ├── eol-pt-15.rst
│ │ │ └── eol-tf-21-24.rst
│ │ └── neuron2.x/
│ │ ├── announce-component-change.rst
│ │ ├── announce-correction-neuron-driver-support-inf1.rst
│ │ ├── announce-deprecation-containers-rtd.rst
│ │ ├── announce-deprecation-nxd-path-trace-api.rst
│ │ ├── announce-deprecation-transformer-flag.rst
│ │ ├── announce-eol-megatron-lm.rst
│ │ ├── announce-eol-python-3-7.rst
│ │ ├── announce-eol-ubuntu-18.rst
│ │ ├── announce-eos-al2.rst
│ │ ├── announce-eos-beta-pytorch-neuroncore-placement-apis.rst
│ │ ├── announce-eos-bf16-vars.rst
│ │ ├── announce-eos-block-dimension-nki.rst
│ │ ├── announce-eos-dlami-ubuntu-22-04.rst
│ │ ├── announce-eos-dlami.rst
│ │ ├── announce-eos-inf1-virtual-environments.rst
│ │ ├── announce-eos-jax-neuronx-nki-call.rst
│ │ ├── announce-eos-megatronlm-2-13.rst
│ │ ├── announce-eos-mllama-checkpoint.rst
│ │ ├── announce-eos-multiframework-dlamis-inf1.rst
│ │ ├── announce-eos-nemo.rst
│ │ ├── announce-eos-neuron-det.rst
│ │ ├── announce-eos-neuron-driver-support-inf1.rst
│ │ ├── announce-eos-neuron-profiler-2.rst
│ │ ├── announce-eos-neuron-profiler-v230.rst
│ │ ├── announce-eos-neuron-profiler.rst
│ │ ├── announce-eos-neurondevice-version.rst
│ │ ├── announce-eos-neurondevice.rst
│ │ ├── announce-eos-nxd-examples.rst
│ │ ├── announce-eos-nxdt-nxd-core-training.rst
│ │ ├── announce-eos-probuf.rst
│ │ ├── announce-eos-pt-versions.rst
│ │ ├── announce-eos-pt2.rst
│ │ ├── announce-eos-python38.rst
│ │ ├── announce-eos-pytorch-1-1-3.rst
│ │ ├── announce-eos-pytorch-1-9.rst
│ │ ├── announce-eos-pytorch-2-1.rst
│ │ ├── announce-eos-pytorch-2-7-2-8-v229.rst
│ │ ├── announce-eos-pytorch-2-7-2-8.rst
│ │ ├── announce-eos-pytorch-profiling-api.rst
│ │ ├── announce-eos-tensorboard-tools.rst
│ │ ├── announce-eos-tensorflow-2-8-9.rst
│ │ ├── announce-eos-tensorflow-inf2.rst
│ │ ├── announce-eos-tensorflow1-x.rst
│ │ ├── announce-eos-torch-neuron.rst
│ │ ├── announce-eos-torch-neuronx-nki-jit.rst
│ │ ├── announce-eos-u20-dlamis.rst
│ │ ├── announce-eos-xla-bf16.rst
│ │ ├── announce-intent-eol-nemo-arg.rst
│ │ ├── announce-intent-eos-opt.rst
│ │ ├── announce-intent-eos-pt-version.rst
│ │ ├── announce-intent-eos-pt2-6.rst
│ │ ├── announce-intent-eos-tensorflow-tutorial-inf.rst
│ │ ├── announce-intent-eos-tnx.rst
│ │ ├── announce-intent-maintenance-tnx.rst
│ │ ├── announce-maintenance-mxnet.rst
│ │ ├── announce-maintenance-nxdi-nxd-core-inference.rst
│ │ ├── announce-maintenance-nxdt-nxd-core-training.rst
│ │ ├── announce-maintenance-tf.rst
│ │ ├── announce-moving-samples.rst
│ │ ├── announce-nki-library-namespace-changes-2-28.rst
│ │ ├── announce-nki-namespace-migration.rst
│ │ ├── announce-no-longer-support-neuron-det.rst
│ │ ├── announce-no-longer-support-nxd-examples.rst
│ │ ├── announce-no-longer-support-pytorch-113.rst
│ │ ├── announce-no-longer-support-pytorch-2-1.rst
│ │ ├── announce-no-longer-support-pytorch-2-7-2-8.rst
│ │ ├── announce-no-longer-support-tensorflow-inf2.rst
│ │ ├── announce-no-longer-support-u20-dlc-dlami.rst
│ │ ├── announce-no-support-al2.rst
│ │ ├── announce-no-support-device-version.rst
│ │ ├── announce-no-support-jax-neuronx-nki-call.rst
│ │ ├── announce-no-support-llama3-2-checkpoint.rst
│ │ ├── announce-no-support-nemo-megatron.rst
│ │ ├── announce-no-support-neurondevice.rst
│ │ ├── announce-no-support-nki-jit-torch.rst
│ │ ├── announce-no-support-tensorboard-plugin.rst
│ │ ├── announce-no-support-tensorflow1-x.rst
│ │ ├── announce-no-support-tensorflow2-10.rst
│ │ ├── announce-no-support-tf-versions.rst
│ │ ├── announce-no-support-torch-neuron-versions.rst
│ │ ├── announce-no-support-ubuntu-20-base.rst
│ │ ├── announce-no-support-vllm-v0.rst
│ │ ├── announce-nxdi-changes.rst
│ │ ├── announce-package-change.rst
│ │ ├── announce-python38-no-longer-support.rst
│ │ ├── announce-transition-pytorch-trainium.rst
│ │ ├── announcement-end-of-support-neuronxcc-nki.rst
│ │ ├── announcement-end-of-support-nxdt-nxd-core.rst
│ │ ├── announcement-end-of-support-parallel-model-trace.rst
│ │ ├── announcement-end-of-support-pytorch-2-6.rst
│ │ ├── announcement-end-of-support-vllm-v0.rst
│ │ ├── announcement-nki-library-kernel-migration.rst
│ │ ├── announcement-nki-library-namespace-changes.rst
│ │ ├── announcement-python-3-9-eol.rst
│ │ ├── dlami-neuron-2.10.rst
│ │ ├── dlami-neuron-2.12.rst
│ │ ├── dlami-pytorch-introduce.rst
│ │ ├── end-of-support-pt2.rst
│ │ ├── github-changes.rst
│ │ ├── gpg-expiration.rst
│ │ ├── neuron-rtd-eol.rst
│ │ ├── neuron2-intro.rst
│ │ ├── neuron230-packages-changes.rst
│ │ ├── neuron250-packages-changes.rst
│ │ ├── release-neuron2.4.rst
│ │ ├── sm-training-dlc-2.9.1.rst
│ │ └── sm-training-trn1-introduce.rst
│ ├── appnotes/
│ │ ├── index.rst
│ │ ├── mxnet-neuron/
│ │ │ └── flex-eg.rst
│ │ ├── neuron-cc/
│ │ │ └── mixed-precision.rst
│ │ ├── neuron1x/
│ │ │ ├── important-neuronx-dkms.txt
│ │ │ └── introducing-libnrt.rst
│ │ ├── neuronx-cc/
│ │ │ └── neuronx-cc-training-mixed-precision.rst
│ │ ├── neuronx-distributed/
│ │ │ ├── introducing-nxd-inference.rst
│ │ │ └── introducing-nxdt-training.rst
│ │ ├── perf/
│ │ │ └── neuron-cc/
│ │ │ ├── parallel-ncgs.rst
│ │ │ └── performance-tuning.rst
│ │ ├── torch-neuron/
│ │ │ ├── bucketing-app-note.rst
│ │ │ ├── index.rst
│ │ │ ├── rcnn-app-note.rst
│ │ │ └── torch-neuron-dataparallel-app-note.rst
│ │ ├── torch-neuronx/
│ │ │ ├── index.rst
│ │ │ ├── introducing-pytorch-2-6.rst
│ │ │ ├── introducing-pytorch-2-7.rst
│ │ │ ├── introducing-pytorch-2-8.rst
│ │ │ ├── introducing-pytorch-2-9.rst
│ │ │ ├── introducing-pytorch-2-x.rst
│ │ │ ├── migration-from-xla-downcast-bf16.rst
│ │ │ ├── torch-neuronx-dataparallel-app-note.rst
│ │ │ └── torch-neuronx-graph-partitioner-app-note.rst
│ │ └── transformers-neuronx/
│ │ └── generative-llm-inference-with-neuron.rst
│ ├── arch/
│ │ ├── glossary.rst
│ │ ├── index.rst
│ │ ├── neuron-features/
│ │ │ ├── custom-c++-operators.rst
│ │ │ ├── data-types.rst
│ │ │ ├── index.rst
│ │ │ ├── logical-neuroncore-config.rst
│ │ │ ├── neuron-caching.rst
│ │ │ ├── neuroncore-batching.rst
│ │ │ ├── neuroncore-pipeline.rst
│ │ │ └── rounding-modes.rst
│ │ └── neuron-hardware/
│ │ ├── inf1-arch.rst
│ │ ├── inf2-arch.rst
│ │ ├── inferentia.rst
│ │ ├── inferentia2.rst
│ │ ├── neuron-core-v1.rst
│ │ ├── neuron-core-v2.rst
│ │ ├── neuron-core-v3.rst
│ │ ├── neuron-core-v4.rst
│ │ ├── trainium.rst
│ │ ├── trainium2.rst
│ │ ├── trainium3.rst
│ │ ├── trn1-arch.rst
│ │ ├── trn2-arch.rst
│ │ └── trn3-arch.rst
│ ├── benchmarks/
│ │ ├── index.rst
│ │ ├── inf1/
│ │ │ ├── data.csv
│ │ │ ├── index.rst
│ │ │ ├── instance_prices.csv
│ │ │ ├── latency_data_encoder.csv
│ │ │ ├── throughput_data_cnn.csv
│ │ │ └── throughput_data_encoder.csv
│ │ ├── inf2/
│ │ │ ├── inf2-performance.rst
│ │ │ ├── inf2_instance_prices.csv
│ │ │ ├── latency_data_decoder.csv
│ │ │ ├── latency_data_encoder.csv
│ │ │ ├── latency_data_encoder_decoder.csv
│ │ │ ├── latency_data_vision.csv
│ │ │ ├── latency_data_vision_cnn.csv
│ │ │ ├── latency_data_vision_dit.csv
│ │ │ ├── latency_data_vision_sd.csv
│ │ │ ├── latency_data_vision_transformers.csv
│ │ │ ├── throughput_data_decoder.csv
│ │ │ ├── throughput_data_encoder.csv
│ │ │ ├── throughput_data_encoder_decoder.csv
│ │ │ ├── throughput_data_vision.csv
│ │ │ ├── throughput_data_vision_cnn.csv
│ │ │ ├── throughput_data_vision_dit.csv
│ │ │ ├── throughput_data_vision_sd.csv
│ │ │ └── throughput_data_vision_transformers.csv
│ │ └── trn1/
│ │ ├── latency_data_decoder.csv
│ │ ├── latency_data_encoder.csv
│ │ ├── latency_data_encoder_decoder.csv
│ │ ├── throughput_data_decoder.csv
│ │ ├── throughput_data_encoder.csv
│ │ ├── throughput_data_encoder_decoder.csv
│ │ ├── training_data_decoder.csv
│ │ ├── training_data_encoder.csv
│ │ ├── training_data_vision_transformers.csv
│ │ ├── trn1-inference-performance.rst
│ │ ├── trn1-training-performance.rst
│ │ ├── trn1_instance_prices.csv
│ │ └── trn1_trn1n_nlp_data.csv
│ ├── beta-participation.rst
│ ├── calculator/
│ │ └── neuron-calculator.rst
│ ├── faq/
│ │ ├── contributing-faq.rst
│ │ ├── index.rst
│ │ ├── inference/
│ │ │ ├── neuron-faq.rst
│ │ │ └── trouble-shooting-faq.rst
│ │ ├── neuron2-intro-faq.rst
│ │ ├── onnx-faq.rst
│ │ ├── roadmap-faq.rst
│ │ └── training/
│ │ └── neuron-training.rst
│ ├── faq.rst
│ ├── index.rst
│ ├── models/
│ │ ├── index.rst
│ │ ├── inference-inf1-samples.rst
│ │ ├── inference-inf2-trn1-samples.rst
│ │ └── training-trn1-samples.rst
│ ├── monitoring-tools.rst
│ ├── news-and-blogs/
│ │ ├── CONTRIBUTING.md
│ │ ├── JIRA-INTEGRATION-DESIGN.md
│ │ ├── README.md
│ │ ├── article-template.yaml
│ │ ├── index.rst
│ │ ├── news-and-blogs.yaml
│ │ └── validate_articles.py
│ ├── oss/
│ │ └── index.rst
│ ├── profiling-tools.rst
│ ├── quick-start/
│ │ ├── _specs/
│ │ │ └── REFACTORING_NOTES.md
│ │ ├── docs-quicklinks.rst
│ │ ├── github-samples.rst
│ │ ├── index.rst
│ │ ├── inference-quickstart.rst
│ │ ├── mxnet-neuron.rst
│ │ ├── tab-inference-tensorflow-neuron.rst
│ │ ├── tensorflow-neuron.rst
│ │ ├── torch-neuron-tab-training.rst
│ │ ├── torch-neuron.rst
│ │ ├── training-quickstart.rst
│ │ └── user-guide-quickstart.rst
│ ├── sdk-policy.rst
│ ├── security.rst
│ ├── troubleshooting.rst
│ ├── what-is-neuron.rst
│ └── whats-new.rst
├── archive/
│ ├── helper-tools/
│ │ ├── index.rst
│ │ ├── tutorial-neuron-check-model.rst
│ │ └── tutorial-neuron-gatherinfo.rst
│ ├── index.rst
│ ├── mxnet-neuron/
│ │ ├── api-compilation-python-api.rst
│ │ ├── api-reference-guide.rst
│ │ ├── api-reference-guide.txt
│ │ ├── developer-guide.rst
│ │ ├── developer-guide.txt
│ │ ├── ec2-then-ec2-devflow.rst
│ │ ├── index.rst
│ │ ├── inference-mxnet-neuron.rst
│ │ ├── inference-mxnet-neuron.txt
│ │ ├── misc-mxnet-neuron.rst
│ │ ├── misc-mxnet-neuron.txt
│ │ ├── mxnet-neuron-setup.rst
│ │ ├── mxnet-neuron-setup.txt
│ │ ├── neo-then-hosting-devflow.rst
│ │ ├── setup/
│ │ │ ├── mxnet-install-prev-al2.rst
│ │ │ ├── mxnet-install-prev-al2023.rst
│ │ │ ├── mxnet-install-prev-u20.rst
│ │ │ ├── mxnet-install-prev-u22.rst
│ │ │ ├── mxnet-install.rst
│ │ │ ├── mxnet-neuron-al2-base-dlami.rst
│ │ │ ├── mxnet-neuron-al2.rst
│ │ │ ├── mxnet-neuron-al2023.rst
│ │ │ ├── mxnet-neuron-ubuntu20-base-dlami.rst
│ │ │ ├── mxnet-neuron-ubuntu20.rst
│ │ │ ├── mxnet-neuron-ubuntu22.rst
│ │ │ ├── mxnet-update-u20.rst
│ │ │ ├── mxnet-update.rst
│ │ │ ├── prev-releases/
│ │ │ │ ├── neuron-1.14.2-mxnet-install.rst
│ │ │ │ ├── neuron-1.15.0-mxnet-install.rst
│ │ │ │ ├── neuron-1.15.1-mxnet-install.rst
│ │ │ │ ├── neuron-1.15.2-mxnet-install.rst
│ │ │ │ ├── neuron-1.16.3-mxnet-install.rst
│ │ │ │ ├── neuron-1.17.2-mxnet-install.rst
│ │ │ │ ├── neuron-1.18.0-mxnet-install.rst
│ │ │ │ └── neuron-1.19.0-mxnet-install.rst
│ │ │ └── setup-inference
│ │ ├── troubleshooting-guide.rst
│ │ └── tutorials/
│ │ ├── mxnet-tutorial-setup.rst
│ │ ├── tutorial-model-serving.rst
│ │ ├── tutorials-mxnet-computervision.rst
│ │ ├── tutorials-mxnet-neuron.rst
│ │ ├── tutorials-mxnet-neuron.txt
│ │ ├── tutorials-mxnet-nlp.rst
│ │ └── tutorials-mxnet-utilizing-neuron-capabilities.rst
│ ├── neuronperf/
│ │ ├── index.rst
│ │ ├── neuronperf_api.rst
│ │ ├── neuronperf_benchmark_guide.rst
│ │ ├── neuronperf_compile_guide.rst
│ │ ├── neuronperf_evaluate_guide.rst
│ │ ├── neuronperf_examples.rst
│ │ ├── neuronperf_faq.rst
│ │ ├── neuronperf_framework_notes.rst
│ │ ├── neuronperf_install.rst
│ │ ├── neuronperf_model_index_guide.rst
│ │ ├── neuronperf_overview.rst
│ │ ├── neuronperf_terminology.rst
│ │ ├── neuronperf_troubleshooting.rst
│ │ ├── rn.rst
│ │ ├── setup.cfg
│ │ ├── setup.py
│ │ ├── test_resnet50_pt.py
│ │ └── test_simple_pt.py
│ ├── src/
│ │ └── benchmark/
│ │ └── pytorch/
│ │ ├── bert-base-cased_benchmark.py
│ │ ├── bert-base-cased_compile.py
│ │ ├── bert-base-uncased_benchmark.py
│ │ ├── bert-base-uncased_compile.py
│ │ ├── distilbert-base-uncased-finetuned-sst-2-english_benchmark.py
│ │ ├── distilbert-base-uncased-finetuned-sst-2-english_compile.py
│ │ ├── distilbert-base-uncased_benchmark.py
│ │ ├── distilbert-base-uncased_compile.py
│ │ ├── distilroberta-base_benchmark.py
│ │ ├── distilroberta-base_compile.py
│ │ ├── hf-google-vit_benchmark.py
│ │ ├── hf-openai-clip_benchmark.py
│ │ ├── hf_pretrained_wav2vec2_conformer_relpos_benchmark.py
│ │ ├── hf_pretrained_wav2vec2_conformer_rope_benchmark.py
│ │ ├── inf2_benchmark.py
│ │ ├── opt_benchmark.py
│ │ ├── perceiver-multimodal_benchmark.py
│ │ ├── perceiver-multimodal_compile.py
│ │ ├── perceiver-vision_benchmark.py
│ │ ├── perceiver-vision_compile.py
│ │ ├── pixart_alpha_benchmark.py
│ │ ├── pixart_sigma_benchmark.py
│ │ ├── resnet50_benchmark.py
│ │ ├── resnet50_compile.py
│ │ ├── resnet_benchmark.py
│ │ ├── resnet_compile.py
│ │ ├── sd2_512_benchmark.py
│ │ ├── sd2_512_compile.py
│ │ ├── sd2_768_benchmark.py
│ │ ├── sd2_768_compile.py
│ │ ├── sd2_inpainting_benchmark.py
│ │ ├── sd2_inpainting_inference.py
│ │ ├── sd_15_512_benchmark.py
│ │ ├── sd_15_512_compile.py
│ │ ├── sd_4x_upscaler_benchmark.py
│ │ ├── sd_4x_upscaler_compile.py
│ │ ├── sdxl_base_1024_benchmark.py
│ │ ├── sdxl_base_1024_compile.py
│ │ ├── sdxl_base_and_refiner_1024_benchmark.py
│ │ ├── sdxl_base_and_refiner_1024_compile.py
│ │ ├── unet_benchmark.py
│ │ ├── unet_compile.py
│ │ ├── vgg_benchmark.py
│ │ └── vgg_compile.py
│ ├── tensorboard/
│ │ └── getting-started-tensorboard-neuron-plugin.rst
│ ├── tensorflow/
│ │ ├── index.rst
│ │ ├── setup-legacy-inf1-tensorflow.rst
│ │ ├── tensorflow-neuron/
│ │ │ ├── additional-examples.rst
│ │ │ ├── additional-examples.txt
│ │ │ ├── api-auto-replication-api.rst
│ │ │ ├── api-compilation-python-api.rst
│ │ │ ├── api-reference-guide.rst
│ │ │ ├── api-reference-guide.txt
│ │ │ ├── api-tfn-analyze-model-api.rst
│ │ │ ├── api-tracing-python-api.rst
│ │ │ ├── dlc-then-ec2-devflow.rst
│ │ │ ├── dlc-then-ecs-devflow.rst
│ │ │ ├── dlc-then-eks-devflow.rst
│ │ │ ├── ec2-then-ec2-devflow.rst
│ │ │ ├── misc-tensorflow-neuron.rst
│ │ │ ├── misc-tensorflow-neuron.txt
│ │ │ ├── neo-then-hosting-devflow.rst
│ │ │ ├── setup/
│ │ │ │ ├── prev-releases/
│ │ │ │ │ ├── neuron-1.14.2-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.15.0-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.15.1-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.15.2-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.16.3-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.17.0-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.17.1-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.17.2-tensorflow-install.rst
│ │ │ │ │ ├── neuron-1.18.0-tensorflow-install.rst
│ │ │ │ │ └── neuron-1.19.0-tensorflow-install.rst
│ │ │ │ ├── tensorflow-install-prev-al2023.rst
│ │ │ │ ├── tensorflow-install-prev-u20.rst
│ │ │ │ ├── tensorflow-install-prev-u22.rst
│ │ │ │ ├── tensorflow-install-prev.rst
│ │ │ │ ├── tensorflow-install.rst
│ │ │ │ ├── tensorflow-update-u20.rst
│ │ │ │ ├── tensorflow-update-u22.rst
│ │ │ │ └── tensorflow-update.rst
│ │ │ ├── tensorflow2-accelerated-ops.rst
│ │ │ ├── tf2_faq.rst
│ │ │ └── tutorials/
│ │ │ ├── bert_demo/
│ │ │ │ ├── bert_demo.rst
│ │ │ │ ├── glue_mrpc_dev.tsv
│ │ │ │ └── mrpc.proto
│ │ │ ├── index.rst
│ │ │ ├── k8s_bert_demo/
│ │ │ │ └── Dockerfile.tfserving_example
│ │ │ ├── tensorflow-tutorial-setup.rst
│ │ │ ├── tutorials-tensorflow-neuron.rst
│ │ │ ├── tutorials-tensorflow-neuron.txt
│ │ │ ├── tutorials-tensorflow-nlp.rst
│ │ │ └── tutorials-tensorflow-utilizing-neuron-capabilities.rst
│ │ ├── tensorflow-neuron-inference.rst
│ │ ├── tensorflow-neuron-inference.txt
│ │ ├── tensorflow-neuronx/
│ │ │ ├── api-reference-guide.rst
│ │ │ ├── api-reference-guide.txt
│ │ │ ├── misc-tensorflow-neuronx.rst
│ │ │ ├── misc-tensorflow-neuronx.txt
│ │ │ ├── setup/
│ │ │ │ ├── index.rst
│ │ │ │ ├── prev-releases/
│ │ │ │ │ ├── neuronx-2.8.0-tensorflow-install.rst
│ │ │ │ │ └── neuronx-2.9.0-tensorflow-install.rst
│ │ │ │ ├── tensorflow-install-prev-al2.rst
│ │ │ │ ├── tensorflow-install-prev-al2023.rst
│ │ │ │ ├── tensorflow-install-prev-u20.rst
│ │ │ │ ├── tensorflow-install-prev-u22.rst
│ │ │ │ ├── tensorflow-neuronx-install.rst
│ │ │ │ ├── tensorflow-update-al2-dlami.rst
│ │ │ │ ├── tensorflow-update-al2.rst
│ │ │ │ ├── tensorflow-update-u20-dlami.rst
│ │ │ │ ├── tensorflow-update-u20.rst
│ │ │ │ └── tensorflow-update-u22.rst
│ │ │ ├── tf-neuronx-auto-replication-api.rst
│ │ │ ├── tfneuronx-python-tracing-api.rst
│ │ │ ├── tfnx-analyze-model-api.rst
│ │ │ └── tutorials/
│ │ │ ├── tutorial-tensorflowx-serving-NeuronRT-Visible-Cores.rst
│ │ │ ├── tutorials-tensorflow-neuronx.rst
│ │ │ └── tutorials-tensorflow-neuronx.txt
│ │ ├── tensorflow-neuronx-inference.rst
│ │ ├── tensorflow-neuronx-inference.txt
│ │ ├── tensorflow-setup.rst
│ │ └── tensorflow-setup.txt
│ ├── torch-neuron/
│ │ ├── additional-examples-inference-torch-neuron.rst
│ │ ├── additional-examples-inference-torch-neuron.txt
│ │ ├── api-compilation-python-api.rst
│ │ ├── api-core-placement.rst
│ │ ├── api-reference-guide-torch-neuron.rst
│ │ ├── api-reference-guide-torch-neuron.txt
│ │ ├── api-torch-neuron-dataparallel-api.rst
│ │ ├── developer-guide-torch-neuron.rst
│ │ ├── developer-guide-torch-neuron.txt
│ │ ├── guides/
│ │ │ ├── core-placement/
│ │ │ │ └── torch-core-placement.rst
│ │ │ └── torch-lstm-support.rst
│ │ ├── index.rst
│ │ ├── inference-torch-neuron.rst
│ │ ├── misc-inference-torch-neuron.rst
│ │ ├── misc-inference-torch-neuron.txt
│ │ ├── placement.py
│ │ ├── setup/
│ │ │ ├── index.rst
│ │ │ ├── prev-releases/
│ │ │ │ ├── neuron-1.14.2-pytorch-install.rst
│ │ │ │ ├── neuron-1.15.0-pytorch-install.rst
│ │ │ │ ├── neuron-1.15.1-pytorch-install.rst
│ │ │ │ ├── neuron-1.15.2-pytorch-install.rst
│ │ │ │ ├── neuron-1.16.1-pytorch-install.rst
│ │ │ │ ├── neuron-1.16.2-pytorch-install.rst
│ │ │ │ ├── neuron-1.16.3-pytorch-install.rst
│ │ │ │ ├── neuron-1.17.2-pytorch-install.rst
│ │ │ │ ├── neuron-1.18.0-pytorch-install.rst
│ │ │ │ ├── neuron-1.19.0-pytorch-install.rst
│ │ │ │ ├── neuron-2.3.0-pytorch-install.rst
│ │ │ │ ├── neuron-2.4.0-pytorch-install.rst
│ │ │ │ └── neuron-2.5.0-pytorch-install.rst
│ │ │ ├── pytorch-install-cxx11.rst
│ │ │ ├── pytorch-install-prev-al2.rst
│ │ │ ├── pytorch-install-prev-al2023.rst
│ │ │ ├── pytorch-install-prev-u20.rst
│ │ │ ├── pytorch-install-prev-u22.rst
│ │ │ ├── pytorch-install-prev.rst
│ │ │ ├── pytorch-install.rst
│ │ │ ├── pytorch-update-al2-dlami.rst
│ │ │ ├── pytorch-update-al2023.rst
│ │ │ ├── pytorch-update-u20-dlami.rst
│ │ │ ├── pytorch-update-u20.rst
│ │ │ ├── pytorch-update-u22.rst
│ │ │ └── pytorch-update.rst
│ │ ├── torch-neuron-dataparallel-example-default.rst
│ │ ├── torch-neuron-dataparallel-example-dim-neq-zero.rst
│ │ ├── torch-neuron-dataparallel-example-disable-dynamic-batching.rst
│ │ ├── torch-neuron-dataparallel-example-dynamic-batching.rst
│ │ ├── torch-neuron-dataparallel-example-specify-ncs.rst
│ │ ├── troubleshooting-guide.rst
│ │ └── tutorials/
│ │ ├── neuroncore_pipeline_pytorch.rst
│ │ ├── pytorch-tutorial-setup.rst
│ │ ├── transformers-marianmt.rst
│ │ ├── tutorial-libtorch.rst
│ │ ├── tutorial-torchserve.rst
│ │ ├── tutorial_source_instructions/
│ │ │ ├── run_libtorch.sh
│ │ │ └── run_torchserve_u20.sh
│ │ ├── tutorials-inference-torch-neuron.rst
│ │ ├── tutorials-inference-torch-neuron.txt
│ │ ├── tutorials-torch-neuron-computervision.rst
│ │ ├── tutorials-torch-neuron-nlp.rst
│ │ └── tutorials-utilizing-neuron-capabilities.rst
│ ├── transformers-neuronx/
│ │ ├── api-reference-guide.rst
│ │ ├── api-reference-guide.txt
│ │ ├── developer-guide.rst
│ │ ├── developer-guide.txt
│ │ ├── index.rst
│ │ ├── setup/
│ │ │ └── index.rst
│ │ ├── transformers-neuronx-api-reference.rst
│ │ ├── transformers-neuronx-developer-guide-for-continuous-batching.rst
│ │ ├── transformers-neuronx-developer-guide.rst
│ │ ├── transformers-neuronx-misc.rst
│ │ ├── transformers-neuronx-misc.txt
│ │ ├── transformers-neuronx-tutorials.rst
│ │ ├── transformers-neuronx-tutorials.txt
│ │ └── transformers-neuronx.txt
│ └── tutorials/
│ ├── finetune_t5.rst
│ ├── finetuning_llama2_7b_ptl.rst
│ ├── gpt3_neuronx_nemo_megatron_pretraining.rst
│ ├── megatron_gpt_pretraining.rst
│ ├── multinode-training-model-profiling.rst
│ ├── nxd-source-code/
│ │ ├── gpt_neox_tp_zero1/
│ │ │ ├── gpt_neox_20b.sh
│ │ │ └── gpt_neox_6_9b.sh
│ │ └── llama_tp_pp_ptl/
│ │ ├── llama_2_13b.sh
│ │ ├── llama_2_70b.sh
│ │ ├── llama_2_7b.sh
│ │ └── llama_tp_pp_ptl_setup.sh
│ ├── ssd300_demo/
│ │ ├── requirements.txt
│ │ ├── ssd300_demo.rst
│ │ ├── ssd300_detection.py
│ │ ├── ssd300_evaluation.py
│ │ ├── ssd300_evaluation_client.py
│ │ └── ssd300_model.py
│ ├── training-gpt-neox-20b.rst
│ ├── training-gpt-neox.rst
│ ├── training_codegen25_7b.rst
│ ├── training_llama2_tp_pp_ptl.rst
│ └── tutorial_source_code/
│ └── t5_finetuning/
│ ├── t5_finetuning_32_worker_training_code.sh
│ ├── t5_finetuning_multi_worker_training_code.sh
│ ├── t5_finetuning_setup_code.sh
│ ├── t5_finetuning_single_worker_training_code.sh
│ └── t5_modify_run_summarization_code.sh
├── audit-report.md
├── build.sh
├── compiler/
│ ├── error-codes/
│ │ ├── EARG001.rst
│ │ ├── EBIR023.rst
│ │ ├── EBVF030.rst
│ │ ├── EHCA005.rst
│ │ ├── EOOM001.rst
│ │ ├── EOOM002.rst
│ │ ├── ESFH002.rst
│ │ ├── ESPP004.rst
│ │ ├── ESPP047.rst
│ │ ├── EUOC002.rst
│ │ ├── EVRF001.rst
│ │ ├── EVRF004.rst
│ │ ├── EVRF005.rst
│ │ ├── EVRF006.rst
│ │ ├── EVRF007.rst
│ │ ├── EVRF009.rst
│ │ ├── EVRF010.rst
│ │ ├── EVRF011.rst
│ │ ├── EVRF013.rst
│ │ ├── EVRF015.rst
│ │ ├── EVRF016.rst
│ │ ├── EVRF017.rst
│ │ ├── EVRF018.rst
│ │ ├── EVRF019.rst
│ │ ├── EVRF022.rst
│ │ ├── EVRF031.rst
│ │ ├── EXSP001.rst
│ │ ├── EXTP004.rst
│ │ └── index.rst
│ ├── index.rst
│ ├── neuron-cc/
│ │ ├── api-reference-guide.rst
│ │ ├── command-line-reference.rst
│ │ ├── developer-guide.rst
│ │ └── faq.rst
│ ├── neuron-cc.rst
│ ├── neuronx-cc/
│ │ ├── api-reference-guide/
│ │ │ └── index.rst
│ │ ├── developer-guide.rst
│ │ ├── faq.rst
│ │ └── how-to-convolution-in-unet.rst
│ └── neuronx-cc.rst
├── conf.py
├── containers/
│ ├── container-deployment-flows.rst
│ ├── container-sm-hosting-devflow.rst
│ ├── developerflows.rst
│ ├── developerflows.txt
│ ├── dlc-then-customize-devflow.rst
│ ├── dlc-then-ec2-devflow.rst
│ ├── dlc-then-ecs-devflow.rst
│ ├── dlc-then-eks-devflow.rst
│ ├── dlc-then-k8s-devflow.rst
│ ├── docker-example/
│ │ ├── Dockerfile.device-plugin
│ │ ├── index.rst
│ │ ├── inference/
│ │ │ ├── Dockerfile-inference
│ │ │ ├── Dockerfile-inference-dlc
│ │ │ ├── Dockerfile-inference-dlc.rst
│ │ │ ├── Dockerfile-libmode
│ │ │ ├── Dockerfile-libmode.rst
│ │ │ ├── Dockerfile-tf-serving.rst
│ │ │ ├── Dockerfile.mxnet-serving
│ │ │ ├── Dockerfile.tf-serving
│ │ │ ├── config-properties.rst
│ │ │ ├── config.properties
│ │ │ ├── dockerd-libmode-entrypoint.rst
│ │ │ ├── dockerd-libmode-entrypoint.sh
│ │ │ ├── torchserve-neuron.rst
│ │ │ └── torchserve-neuron.sh
│ │ ├── training/
│ │ │ ├── Dockerfile-training-dlc
│ │ │ ├── Dockerfile-trainium-dlc.rst
│ │ │ ├── mlp.rst
│ │ │ ├── mlp_train.py
│ │ │ └── model.py
│ │ └── v1/
│ │ └── inference/
│ │ ├── Dockerfile-app-rt-diff.rst
│ │ ├── Dockerfile-app-rt-same.rst
│ │ ├── Dockerfile-neuron-rtd.rst
│ │ ├── Dockerfile-torch-neuron.rst
│ │ ├── Dockerfile.app-rt-diff
│ │ ├── Dockerfile.neuron-rtd
│ │ ├── Dockerfile.torch-neuron
│ │ ├── dockerd-entrypoint-app-rt-same.rst
│ │ └── dockerd-entrypoint.sh
│ ├── ec2-then-ec2-devflow.rst
│ ├── ec2.rst
│ ├── faq-troubleshooting-releasenote.rst
│ ├── faq.rst
│ ├── files/
│ │ ├── index-dra.rst
│ │ ├── manifests/
│ │ │ ├── clusterrole.yaml
│ │ │ ├── clusterrolebinding.yaml
│ │ │ ├── daemonset.yaml
│ │ │ ├── deviceclass.yaml
│ │ │ ├── namespace.yaml
│ │ │ └── serviceaccount.yaml
│ │ ├── scripts/
│ │ │ └── install-dra-driver.sh
│ │ └── specs/
│ │ ├── 1x4-connected-devices.yaml
│ │ ├── 2-node-inference-us.yaml
│ │ ├── 4-node-inference-us.yaml
│ │ ├── all-devices.yaml
│ │ ├── lnc-setting-trn2.yaml
│ │ ├── specific-driver-version.yaml
│ │ └── us-and-lnc-config.yaml
│ ├── get-started/
│ │ ├── quickstart-configure-deploy-dlc.rst
│ │ └── quickstart-pytorch-inference-dlc.rst
│ ├── getting-started.rst
│ ├── how-to/
│ │ └── how-to-ultraserver.rst
│ ├── index.rst
│ ├── k8.rst
│ ├── kubernetes-getting-started.rst
│ ├── locate-neuron-dlc-image.rst
│ ├── neo-then-hosting-devflow.rst
│ ├── neuron-dra.rst
│ ├── neuron-plugins.rst
│ ├── neuron_dlc_images.csv
│ ├── troubleshooting.rst
│ ├── tutorial-docker-runtime1.0.rst
│ ├── tutorials/
│ │ ├── build-run-neuron-container.rst
│ │ ├── inference/
│ │ │ ├── index.rst
│ │ │ ├── index.txt
│ │ │ ├── k8s_rn50_demo.rst
│ │ │ └── tutorial-infer.rst
│ │ ├── k8s-default-scheduler.rst
│ │ ├── k8s-multiple-scheduler.rst
│ │ ├── k8s-neuron-device-plugin.rst
│ │ ├── k8s-neuron-helm-chart.rst
│ │ ├── k8s-neuron-monitor.rst
│ │ ├── k8s-neuron-problem-detector-and-recovery-irsa.rst
│ │ ├── k8s-neuron-problem-detector-and-recovery.rst
│ │ ├── k8s-neuron-scheduler-flow.rst
│ │ ├── k8s-neuron-scheduler.rst
│ │ ├── k8s-prerequisite.rst
│ │ ├── k8s-setup.rst
│ │ ├── training/
│ │ │ ├── index.rst
│ │ │ ├── index.txt
│ │ │ ├── k8s_mlp_train_demo.rst
│ │ │ └── tutorial-training.rst
│ │ ├── tutorial-docker-env-setup.rst
│ │ └── tutorial-oci-hook.rst
│ └── tutorials.rst
├── devflows/
│ ├── aws-batch-flows.rst
│ ├── aws-batch-flows.txt
│ ├── dlc-then-customize-devflow.rst
│ ├── ec2-flows.rst
│ ├── ec2-flows.txt
│ ├── ecs-flows.rst
│ ├── eks-flows.rst
│ ├── index.rst
│ ├── inference/
│ │ ├── aws-batch-flows.rst
│ │ ├── aws-batch-flows.txt
│ │ ├── byoc-hosting-devflow-inf2.rst
│ │ ├── byoc-hosting-devflow.rst
│ │ ├── container-sm-hosting-devflow.rst
│ │ ├── dev-flows.rst
│ │ ├── dlc-then-ec2-devflow.rst
│ │ ├── dlc-then-ecs-devflow.rst
│ │ ├── dlc-then-eks-devflow.rst
│ │ ├── dlc-then-k8s-devflow.rst
│ │ ├── ec2-flows.rst
│ │ ├── ec2-flows.txt
│ │ ├── ec2-then-ec2-devflow-inf2.rst
│ │ ├── ec2-then-ec2-devflow.rst
│ │ ├── env-setup-text.rst
│ │ ├── neo-then-hosting-devflow.rst
│ │ ├── parallelcluster-flows.rst
│ │ ├── parallelcluster-flows.txt
│ │ ├── sagemaker-flows.rst
│ │ └── sagemaker-flows.txt
│ ├── parallelcluster-flows.rst
│ ├── parallelcluster-flows.txt
│ ├── plugins/
│ │ ├── npd-ecs-flows.rst
│ │ └── npd-ecs-flows.txt
│ ├── sagemaker-flows.rst
│ ├── setup/
│ │ ├── ecs-flows.rst
│ │ ├── ecs-flows.txt
│ │ ├── eks-flows.rst
│ │ └── eks-flows.txt
│ ├── third-party-solutions.rst
│ └── training/
│ ├── aws-batch-flows.rst
│ ├── aws-batch-flows.txt
│ ├── batch/
│ │ └── batch-training.rst
│ ├── dlc-then-ecs-devflow.rst
│ ├── ec2/
│ │ └── ec2-training.rst
│ ├── ec2-flows.rst
│ ├── ec2-flows.txt
│ ├── parallelcluster/
│ │ └── parallelcluster-training.rst
│ ├── parallelcluster-flows.rst
│ ├── parallelcluster-flows.txt
│ ├── sagemaker-flows.rst
│ ├── sagemaker-flows.txt
│ └── sm-devflow/
│ └── sm-training-devflow.rst
├── dlami/
│ └── index.rst
├── frameworks/
│ ├── index.rst
│ ├── jax/
│ │ ├── api-reference-guide/
│ │ │ ├── index.rst
│ │ │ └── neuron-envvars.rst
│ │ ├── index.rst
│ │ └── setup/
│ │ ├── jax-neuronx-known-issues.rst
│ │ └── jax-setup.rst
│ └── torch/
│ ├── about/
│ │ └── index.rst
│ ├── guide-torch-neuron-vs-torch-neuronx-inference.rst
│ ├── index.rst
│ ├── inference-torch-neuronx.rst
│ ├── pytorch-native-overview.rst
│ ├── torch-neuronx/
│ │ ├── additional-examples-inference-torch-neuronx.rst
│ │ ├── additional-examples-training.rst
│ │ ├── api-reference-guide/
│ │ │ ├── inference/
│ │ │ │ ├── api-torch-neuronx-analyze.rst
│ │ │ │ ├── api-torch-neuronx-async-lazy-load.rst
│ │ │ │ ├── api-torch-neuronx-core-placement.rst
│ │ │ │ ├── api-torch-neuronx-data-parallel.rst
│ │ │ │ ├── api-torch-neuronx-replace-weights.rst
│ │ │ │ ├── api-torch-neuronx-trace.rst
│ │ │ │ └── inference-api-guide-torch-neuronx.rst
│ │ │ ├── torch-neuronx-profiling-api.rst
│ │ │ └── training/
│ │ │ ├── index.rst
│ │ │ ├── pytorch-neuron-parallel-compile.rst
│ │ │ └── torch-neuron-envvars.rst
│ │ ├── misc-inference-torch-neuronx.rst
│ │ ├── misc-training.rst
│ │ ├── programming-guide/
│ │ │ ├── inference/
│ │ │ │ ├── autobucketing-dev-guide.rst
│ │ │ │ ├── core-placement.rst
│ │ │ │ ├── index.rst
│ │ │ │ └── trace-vs-xla-lazytensor.rst
│ │ │ ├── torch-neuronx-profiling-dev-guide.rst
│ │ │ └── training/
│ │ │ ├── index.rst
│ │ │ ├── pytorch-neuron-debug.rst
│ │ │ └── pytorch-neuron-programming-guide.rst
│ │ ├── pytorch-neuron-supported-operators.rst
│ │ ├── setup/
│ │ │ ├── install-templates/
│ │ │ │ └── pytorch-dev-install.txt
│ │ │ ├── note-setup-general.rst
│ │ │ ├── prev-releases/
│ │ │ │ ├── neuronx-2.7.0-pytorch-install.rst
│ │ │ │ ├── neuronx-2.8.0-pytorch-install.rst
│ │ │ │ └── neuronx-2.9.0-pytorch-install.rst
│ │ │ ├── pytorch-install-prev-al2.rst
│ │ │ ├── pytorch-install-prev-al2023.rst
│ │ │ ├── pytorch-install-prev-u20.rst
│ │ │ ├── pytorch-install-prev-u22.rst
│ │ │ ├── pytorch-install-prev-u24.rst
│ │ │ ├── pytorch-install.rst
│ │ │ ├── pytorch-neuronx-install-cxx11.rst
│ │ │ ├── pytorch-update-al2-dlami.rst
│ │ │ ├── pytorch-update-al2.rst
│ │ │ ├── pytorch-update-al2023.rst
│ │ │ ├── pytorch-update-u20-dlami.rst
│ │ │ ├── pytorch-update-u20.rst
│ │ │ ├── pytorch-update-u22.rst
│ │ │ └── pytorch-update-u24.rst
│ │ ├── setup-trn1-multi-node-execution.rst
│ │ ├── torch-neuronx-dataparallel-example-default.rst
│ │ ├── torch-neuronx-dataparallel-example-dim-neq-zero.rst
│ │ ├── torch-neuronx-dataparallel-example-disable-dynamic-batching.rst
│ │ ├── torch-neuronx-dataparallel-example-dynamic-batching.rst
│ │ ├── torch-neuronx-dataparallel-example-specify-ncs.rst
│ │ ├── training-troubleshooting.rst
│ │ └── tutorials/
│ │ ├── inference/
│ │ │ ├── tutorial-torchserve-neuronx.rst
│ │ │ └── tutorials-torch-neuronx.rst
│ │ ├── note-performance.txt
│ │ └── training/
│ │ ├── analyze_for_training.rst
│ │ ├── bert.rst
│ │ ├── finetune_hftrainer.rst
│ │ ├── mlp.rst
│ │ ├── tutorial_source_code/
│ │ │ ├── analyze_training/
│ │ │ │ └── analyze_training_code.sh
│ │ │ ├── bert_mrpc_finetuning/
│ │ │ │ ├── bert_mrpc_finetuning_converted_checkpoint_training.sh
│ │ │ │ ├── bert_mrpc_finetuning_multi_worker_training_code.sh
│ │ │ │ ├── bert_mrpc_finetuning_setup_code.sh
│ │ │ │ └── bert_mrpc_finetuning_single_worker_training.sh
│ │ │ ├── bert_training/
│ │ │ │ ├── bert_amp_training_code.sh
│ │ │ │ ├── bert_lamb_bf16_training_code.sh
│ │ │ │ ├── bert_lamb_training_code.sh
│ │ │ │ ├── bert_phase2_training_code.sh
│ │ │ │ ├── bert_precompilation_code.sh
│ │ │ │ ├── bert_setup_code.sh
│ │ │ │ ├── bert_setup_code_ph2.sh
│ │ │ │ └── bert_training_code.sh
│ │ │ ├── multi_layer_perceptron_training/
│ │ │ │ └── multi_layer_perceptron_training_code.sh
│ │ │ └── zero1_training/
│ │ │ └── zero1_single_node_training_code.sh
│ │ ├── tutorials-training-torch-neuronx.rst
│ │ └── zero1_gpt2.rst
│ ├── torch-setup.rst
│ └── training-torch-neuronx.rst
├── general/
│ └── faq.rst
├── includes/
│ └── setup/
│ ├── select-framework-note.txt
│ ├── tab-inference-mxnet-neuron-al2.txt
│ ├── tab-inference-mxnet-neuron-al2023.txt
│ ├── tab-inference-mxnet-neuron-u20.txt
│ ├── tab-inference-mxnet-neuron-u22.txt
│ ├── tab-inference-mxnet-neuron.txt
│ ├── tab-inference-tensorflow-neuron-al2.txt
│ ├── tab-inference-tensorflow-neuron-al2023.txt
│ ├── tab-inference-tensorflow-neuron-u20.txt
│ ├── tab-inference-tensorflow-neuron-u22.txt
│ ├── tab-inference-tensorflow-neuronx-al2.txt
│ ├── tab-inference-tensorflow-neuronx-al2023.txt
│ ├── tab-inference-tensorflow-neuronx-u20.txt
│ ├── tab-inference-tensorflow-neuronx-u22.txt
│ ├── tab-inference-torch-neuron-al2.txt
│ ├── tab-inference-torch-neuron-al2023.txt
│ ├── tab-inference-torch-neuron-u20.txt
│ ├── tab-inference-torch-neuron-u22.txt
│ ├── tab-inference-torch-neuron.txt
│ ├── tab-inference-torch-neuronx-al2.txt
│ ├── tab-inference-torch-neuronx-al2023.txt
│ ├── tab-inference-torch-neuronx-u20.txt
│ ├── tab-inference-torch-neuronx-u22.txt
│ └── tab-inference-torch-neuronx-u24.txt
├── index.rst
├── info/
│ └── exclude
├── libraries/
│ ├── index.rst
│ ├── nemo-megatron/
│ │ └── index.rst
│ ├── neuronx-distributed/
│ │ ├── activation_memory_reduction.rst
│ │ ├── activation_memory_reduction_developer_guide.rst
│ │ ├── api-reference-guide-inference.rst
│ │ ├── api-reference-guide-training.rst
│ │ ├── api-reference-guide.rst
│ │ ├── api-reference-guide.txt
│ │ ├── api_guide.rst
│ │ ├── app_notes.rst
│ │ ├── app_notes.txt
│ │ ├── context_parallelism_overview.rst
│ │ ├── developer-guide-inference.rst
│ │ ├── developer-guide-inference.txt
│ │ ├── developer-guide-training.rst
│ │ ├── developer-guide-training.txt
│ │ ├── developer-guide.rst
│ │ ├── developer-guide.txt
│ │ ├── index-inference.rst
│ │ ├── index-training.rst
│ │ ├── lora_finetune_developer_guide.rst
│ │ ├── model_builder_v2_api_reference.rst
│ │ ├── model_optimizer_wrapper_developer_guide.rst
│ │ ├── neuronx-distributed-misc.rst
│ │ ├── neuronx-distributed-misc.txt
│ │ ├── neuronx_distributed_inference_developer_guide.rst
│ │ ├── pipeline_parallelism_overview.rst
│ │ ├── pp_developer_guide.rst
│ │ ├── ptl_developer_guide.rst
│ │ ├── save_load_developer_guide.rst
│ │ ├── setup/
│ │ │ ├── index.rst
│ │ │ └── index.txt
│ │ ├── standard_mixed_precision.rst
│ │ ├── tensor_parallelism_overview.rst
│ │ ├── tp_developer_guide.rst
│ │ └── tutorials/
│ │ ├── finetune_llama3_8b_ptl_lora.rst
│ │ ├── index.rst
│ │ ├── index.txt
│ │ ├── inference.rst
│ │ ├── inference_tutorials.rst
│ │ ├── neuronx_distributed_tutorials.txt
│ │ ├── nxd-source-code/
│ │ │ ├── llama_tp_pp/
│ │ │ │ ├── llama_2_13b.sh
│ │ │ │ ├── llama_2_70b.sh
│ │ │ │ ├── llama_31_70b.sh
│ │ │ │ ├── llama_3_70b.sh
│ │ │ │ └── llama_tp_pp_setup.sh
│ │ │ └── llama_tp_zero1/
│ │ │ ├── llama_2_7b.sh
│ │ │ ├── llama_31_8b.sh
│ │ │ ├── llama_3_8b.sh
│ │ │ └── llama_tp_zero1_setup.sh
│ │ ├── nxd_inference_tutorials.txt
│ │ ├── nxd_training_tutorials.txt
│ │ ├── training.rst
│ │ ├── training_llama_tp_pp.rst
│ │ ├── training_llama_tp_zero1.rst
│ │ └── training_tutorials.rst
│ ├── nxd-inference/
│ │ ├── _templates/
│ │ │ ├── model_card.jinja.rst
│ │ │ └── model_card_qwen3.jinja.rst
│ │ ├── api-guides/
│ │ │ ├── api-guide.rst
│ │ │ ├── api-guide.txt
│ │ │ └── index.rst
│ │ ├── app-notes/
│ │ │ ├── app_notes.txt
│ │ │ ├── index.rst
│ │ │ └── parallelism.rst
│ │ ├── developer_guides/
│ │ │ ├── accuracy-eval-with-datasets.rst
│ │ │ ├── custom-quantization.rst
│ │ │ ├── disaggregated-inference.rst
│ │ │ ├── feature-guide.rst
│ │ │ ├── how-to-use-fpem.rst
│ │ │ ├── index.rst
│ │ │ ├── llm-inference-benchmarking-guide.rst
│ │ │ ├── migrate-from-tnx-to-nxdi.rst
│ │ │ ├── model-reference.rst
│ │ │ ├── moe-arch-deep-dive.rst
│ │ │ ├── nxd-examples-migration-guide.rst
│ │ │ ├── onboarding-models.rst
│ │ │ ├── performance-cli-params.rst
│ │ │ ├── vllm-user-guide-v1.rst
│ │ │ ├── vllm-user-guide.rst
│ │ │ ├── weights-sharding-guide.rst
│ │ │ └── writing-tests.rst
│ │ ├── examples/
│ │ │ └── vllm_client.py
│ │ ├── index.rst
│ │ ├── misc/
│ │ │ ├── index.rst
│ │ │ ├── misc.txt
│ │ │ └── nxdi-troubleshooting.rst
│ │ ├── models/
│ │ │ ├── index.rst
│ │ │ ├── llama3/
│ │ │ │ ├── data/
│ │ │ │ │ └── card_llama33_70b.yml
│ │ │ │ └── llama_33_70b.rst
│ │ │ ├── models.txt
│ │ │ └── qwen3/
│ │ │ ├── data/
│ │ │ │ └── card_qwen3_moe_235b.yml
│ │ │ └── qwen3_moe_235b.rst
│ │ ├── neuron-inference-overview.rst
│ │ ├── nxdi-setup.rst
│ │ ├── overview-index.rst
│ │ ├── setup.txt
│ │ ├── tutorials/
│ │ │ ├── disaggregated-inference-tutorial-1p1d.rst
│ │ │ ├── disaggregated-inference-tutorial.rst
│ │ │ ├── flux-inference-tutorial.ipynb
│ │ │ ├── flux-inpainting-inference-tutorial.ipynb
│ │ │ ├── generating-results-with-performance-cli.ipynb
│ │ │ ├── index.rst
│ │ │ ├── llama4-tutorial-v0.ipynb
│ │ │ ├── llama4-tutorial.ipynb
│ │ │ ├── llama405b_perf_comparison.csv
│ │ │ ├── llama70b_apc_perf_comparison.csv
│ │ │ ├── llama70b_perf_comparison.csv
│ │ │ ├── modules_to_not_convert.json
│ │ │ ├── pixtral-tutorial.ipynb
│ │ │ ├── qwen2-vl-tutorial.ipynb
│ │ │ ├── qwen3-moe-tutorial.ipynb
│ │ │ ├── qwen3-vl-tutorial.ipynb
│ │ │ ├── sd-inference-tutorial.rst
│ │ │ ├── trn1-llama3.1-70b-instruct-accuracy-eval-tutorial.ipynb
│ │ │ ├── trn2-llama3.1-405b-speculative-tutorial.rst
│ │ │ ├── trn2-llama3.1-405b-tutorial.rst
│ │ │ ├── trn2-llama3.1-8b-multi-lora-tutorial.ipynb
│ │ │ ├── trn2-llama3.3-70b-apc-tutorial.ipynb
│ │ │ ├── trn2-llama3.3-70b-dp-tutorial.ipynb
│ │ │ ├── trn2-llama3.3-70b-fp8.rst
│ │ │ ├── trn2-llama3.3-70b-tutorial.rst
│ │ │ └── trn3-gpt-oss-120b-tutorial.rst
│ │ └── vllm/
│ │ ├── index.rst
│ │ ├── quickstart-vllm-offline-serving.rst
│ │ └── quickstart-vllm-online-serving.rst
│ ├── nxd-training/
│ │ ├── api-guide.txt
│ │ ├── api-reference-guide.rst
│ │ ├── app_notes/
│ │ │ ├── nxd-training-amr-appnote.rst
│ │ │ ├── nxd-training-cp-appnote.rst
│ │ │ ├── nxd-training-pp-appnote.rst
│ │ │ └── nxd-training-tp-appnote.rst
│ │ ├── app_notes.rst
│ │ ├── app_notes.txt
│ │ ├── developer-guide.rst
│ │ ├── developer_guides/
│ │ │ ├── cpu_mode_developer_guide.rst
│ │ │ ├── dev-guide.txt
│ │ │ ├── index.rst
│ │ │ ├── migration_nemo_nxdt.rst
│ │ │ ├── migration_nnm_nxdt.rst
│ │ │ ├── nemo_nxdt_mapping.csv
│ │ │ ├── new_dataloader_guide.rst
│ │ │ ├── new_model_guide.rst
│ │ │ ├── nnm_nxdt_mapping.csv
│ │ │ └── optimizer_lr_scheduler_flow.rst
│ │ ├── general/
│ │ │ ├── config_overview.rst
│ │ │ ├── features.rst
│ │ │ ├── installation_guide.rst
│ │ │ ├── known-issues.txt
│ │ │ └── known_issues.rst
│ │ ├── index.rst
│ │ ├── misc.rst
│ │ ├── misc.txt
│ │ ├── overview.rst
│ │ ├── overview.txt
│ │ ├── setup.txt
│ │ └── tutorials/
│ │ ├── checkpoint_conversion.rst
│ │ ├── hf_llama3_70B_pretraining.rst
│ │ ├── hf_llama3_8B_DPO_ORPO.rst
│ │ ├── hf_llama3_8B_SFT.rst
│ │ ├── hf_llama3_8B_SFT_LORA.rst
│ │ ├── hf_llama3_8B_pretraining.rst
│ │ ├── index.rst
│ │ ├── megatron_gpt_pretraining.rst
│ │ └── tutorials.txt
│ └── transformers-neuronx/
│ └── index.rst
├── llms.txt
├── neuron-customops/
│ ├── api-reference-guide/
│ │ ├── api-reference-guide.rst
│ │ └── custom-ops-ref-guide.rst
│ ├── customops-intro.txt
│ ├── index.rst
│ ├── misc-customops.rst
│ ├── programming-guide/
│ │ ├── custom-c++-operators-devguide.rst
│ │ └── programming-guide.rst
│ └── tutorials/
│ ├── customop-mlp-perf-opt.rst
│ ├── customop-mlp-training.rst
│ ├── tutorial_source_code/
│ │ ├── custom_c_mlp_training/
│ │ │ └── custom_c_mlp_training_code.sh
│ │ └── custom_c_perf_optimization/
│ │ └── custom_c_perf_optimization_code.sh
│ └── tutorials.rst
├── neuron-runtime/
│ ├── about/
│ │ ├── collectives.rst
│ │ ├── core-dump.rst
│ │ └── index.rst
│ ├── api/
│ │ ├── debug-stream-api.rst
│ │ ├── index.rst
│ │ ├── ndebug_stream.rst
│ │ ├── ndl.rst
│ │ ├── nec.rst
│ │ ├── neuron_driver_shared.rst
│ │ ├── neuron_driver_shared_tensor_batch_op.rst
│ │ ├── neuron_ds.rst
│ │ ├── nrt-async-api-best-practices.rst
│ │ ├── nrt-async-api-examples.rst
│ │ ├── nrt-async-api-overview.rst
│ │ ├── nrt.rst
│ │ ├── nrt_async.rst
│ │ ├── nrt_async_sendrecv.rst
│ │ ├── nrt_experimental.rst
│ │ ├── nrt_profile.rst
│ │ ├── nrt_status.rst
│ │ ├── nrt_sys_trace.rst
│ │ └── nrt_version.rst
│ ├── configuration-guide.rst
│ ├── explore/
│ │ ├── compute-comm-overlap.rst
│ │ ├── core-dump-deep-dive.rst
│ │ ├── device-memory.rst
│ │ ├── direct-hbm-tensor-alloc.rst
│ │ ├── index.rst
│ │ ├── internode-collective-comm.rst
│ │ ├── intranode-collective-comm.rst
│ │ ├── runtime-performance-tips.rst
│ │ └── work-with-neff-files.rst
│ ├── faq.rst
│ ├── index.rst
│ ├── nrt-configurable-parameters.rst
│ ├── nrt-developer-guide.rst
│ ├── nrt-troubleshoot.rst
│ └── rn.rst
├── nki/
│ ├── _ext/
│ │ └── nki_directives.py
│ ├── _templates/
│ │ ├── nki-custom-class-attr-only-template.rst
│ │ └── nki-custom-class-template.rst
│ ├── api/
│ │ ├── index.rst
│ │ ├── nki/
│ │ │ ├── __init__.py
│ │ │ ├── collectives/
│ │ │ │ └── __init__.py
│ │ │ ├── isa/
│ │ │ │ └── __init__.py
│ │ │ └── language/
│ │ │ └── __init__.py
│ │ ├── nki.api.shared.rst
│ │ ├── nki.collectives.rst
│ │ ├── nki.isa.rst
│ │ ├── nki.isa.rst.bak
│ │ ├── nki.language.rst
│ │ ├── nki.language.tile_size.rst
│ │ ├── nki.rst
│ │ └── nki.simulate.rst
│ ├── deep-dives/
│ │ ├── index.rst
│ │ ├── mxfp-matmul.rst
│ │ ├── nki-aps.rst
│ │ ├── nki-compiler.rst
│ │ ├── nki-dge.rst
│ │ ├── nki-dma-bandwidth-guide.rst
│ │ ├── nki-dynamic-loops.rst
│ │ ├── nki_perf_guide.rst
│ │ └── src/
│ │ └── mxfp-matmul/
│ │ ├── mx_cpu_utils.py
│ │ ├── mx_kernel_utils.py
│ │ ├── mx_kernels.py
│ │ └── mx_toplevel.py
│ ├── examples/
│ │ ├── average_pool2d/
│ │ │ ├── average_pool2d_jax.py
│ │ │ ├── average_pool2d_nki_kernels.py
│ │ │ └── average_pool2d_torch.py
│ │ ├── fused_mamba/
│ │ │ ├── mamba_nki_kernels.py
│ │ │ └── mamba_torch.py
│ │ ├── getting_started_baremetal.py
│ │ ├── getting_started_jax.py
│ │ ├── getting_started_torch.py
│ │ ├── index-case-1.py
│ │ ├── index-case-3.py
│ │ ├── layout-dynamic-loop.py
│ │ ├── layout-loop.py
│ │ ├── layout-pass.py
│ │ ├── layout-violation.py
│ │ ├── matrix_multiplication/
│ │ │ ├── matrix_multiplication_nki_kernels.py
│ │ │ └── matrix_multiplication_torch.py
│ │ ├── simulate/
│ │ │ └── nki_simulate_example.py
│ │ ├── tensor_addition/
│ │ │ └── tensor_addition_nki_kernels.py
│ │ └── transpose2d/
│ │ ├── transpose2d_jax.py
│ │ ├── transpose2d_nki_kernels.py
│ │ └── transpose2d_torch.py
│ ├── get-started/
│ │ ├── about/
│ │ │ ├── data-representation-overview.rst
│ │ │ ├── index.rst
│ │ │ ├── indexing-overview.rst
│ │ │ ├── lnc.rst
│ │ │ ├── memory-hierarchy-overview.rst
│ │ │ ├── nki-dma-overview.rst
│ │ │ └── tiling-overview.rst
│ │ ├── index.rst
│ │ ├── nki-language-guide.rst
│ │ ├── quickstart-implement-run-kernel.rst
│ │ └── setup-env.rst
│ ├── guides/
│ │ ├── architecture/
│ │ │ ├── index.rst
│ │ │ ├── trainium2_arch.rst
│ │ │ ├── trainium3_arch.rst
│ │ │ └── trainium_inferentia2_arch.rst
│ │ ├── framework_custom_op.rst
│ │ ├── how-to-scheduling-apis.rst
│ │ ├── index.rst
│ │ ├── nki_simulator.rst
│ │ ├── tutorials/
│ │ │ ├── average_pool2d.rst
│ │ │ ├── fused_mamba.rst
│ │ │ ├── index.rst
│ │ │ ├── kernel-optimization.rst
│ │ │ ├── matrix_multiplication.rst
│ │ │ └── transpose2d.rst
│ │ └── use-neuron-profile.rst
│ ├── index.rst
│ ├── library/
│ │ ├── about/
│ │ │ └── index.rst
│ │ ├── api/
│ │ │ ├── attention-block-tkg.rst
│ │ │ ├── attention-cte.rst
│ │ │ ├── attention-tkg.rst
│ │ │ ├── blockwise-mm-backward.rst
│ │ │ ├── conv1d.rst
│ │ │ ├── cross-entropy.rst
│ │ │ ├── cumsum.rst
│ │ │ ├── depthwise-conv1d.rst
│ │ │ ├── dynamic-elementwise-add.rst
│ │ │ ├── fg-allgather.rst
│ │ │ ├── fgcc.rst
│ │ │ ├── find-nonzero-indices.rst
│ │ │ ├── index.rst
│ │ │ ├── mlp.rst
│ │ │ ├── moe-cte.rst
│ │ │ ├── moe-tkg.rst
│ │ │ ├── output-projection-cte.rst
│ │ │ ├── output-projection-tkg.rst
│ │ │ ├── qkv.rst
│ │ │ ├── rmsnorm-quant.rst
│ │ │ ├── rope.rst
│ │ │ ├── router-topk.rst
│ │ │ ├── sb2sb-allgather.rst
│ │ │ ├── topk-reduce.rst
│ │ │ └── transformer-tkg.rst
│ │ ├── index.rst
│ │ ├── kernel-utils/
│ │ │ ├── allocator.rst
│ │ │ ├── index.rst
│ │ │ └── tensor-view.rst
│ │ └── specs/
│ │ ├── design-rmsnorm-quant.rst
│ │ └── index.rst
│ ├── migration/
│ │ ├── index.rst
│ │ ├── nki-0-3-0-update-guide.rst
│ │ ├── nki-beta2-migration-guide.rst
│ │ └── nki_block_dimension_migration_guide.rst
│ ├── nki_faq.rst
│ ├── scripts/
│ │ ├── markdown2rst.py
│ │ └── requirements.txt
│ └── test/
│ ├── test_nki_isa_activation.py
│ ├── test_nki_isa_affine_select.py
│ ├── test_nki_isa_bn_stats.py
│ ├── test_nki_isa_copypredicated.py
│ ├── test_nki_isa_dma_copy.py
│ ├── test_nki_isa_dma_transpose.py
│ ├── test_nki_isa_dropout.py
│ ├── test_nki_isa_iota.py
│ ├── test_nki_isa_local_gather.py
│ ├── test_nki_isa_max8.py
│ ├── test_nki_isa_memset.py
│ ├── test_nki_isa_nc_find_index8.py
│ ├── test_nki_isa_nc_match_replace8.py
│ ├── test_nki_isa_nc_matmul.py
│ ├── test_nki_isa_nc_stream_shuffle.py
│ ├── test_nki_isa_nc_transpose.py
│ ├── test_nki_isa_partition_reduce.py
│ ├── test_nki_isa_range_select.py
│ ├── test_nki_isa_reciprocal.py
│ ├── test_nki_isa_reduce.py
│ ├── test_nki_isa_select_reduce.py
│ ├── test_nki_isa_sequence_bounds.py
│ ├── test_nki_isa_tensor_copy.py
│ ├── test_nki_isa_tensor_scalar.py
│ ├── test_nki_isa_tensor_scalar_cumulative.py
│ ├── test_nki_isa_tensor_tensor.py
│ ├── test_nki_isa_tensor_tensor_scan.py
│ ├── test_nki_mask.py
│ ├── test_nki_memory_semantics.py
│ ├── test_nki_nl_add.py
│ ├── test_nki_nl_atomic_rmw.py
│ ├── test_nki_nl_broadcast.py
│ ├── test_nki_nl_dslice.py
│ ├── test_nki_nl_gather_flattened.py
│ ├── test_nki_nl_load_store.py
│ ├── test_nki_nl_load_store_indirect.py
│ ├── test_nki_nl_load_transpose2d.py
│ ├── test_nki_nl_mgrid.py
│ ├── test_nki_simulate_kernel.py
│ ├── test_nki_spmd_grid.py
│ ├── test_psum_modulo_alloc.py
│ └── test_sbuf_modulo_alloc.py
├── release-notes/
│ ├── 2.29.0.rst
│ ├── archive/
│ │ ├── customcxxps/
│ │ │ ├── gpsimd-customop-lib.rst
│ │ │ └── gpsimd-tools.rst
│ │ ├── index.rst
│ │ ├── libneuronxla.rst
│ │ ├── mxnet-neuron.rst
│ │ ├── nemo/
│ │ │ ├── index.rst
│ │ │ └── neuronx-nemo.rst
│ │ ├── neuron-cc/
│ │ │ ├── neuron-cc-ops/
│ │ │ │ ├── index.rst
│ │ │ │ ├── neuron-cc-ops-mxnet.rst
│ │ │ │ ├── neuron-cc-ops-pytorch.rst
│ │ │ │ ├── neuron-cc-ops-tensorflow.rst
│ │ │ │ └── neuron-cc-ops-xla.rst
│ │ │ └── neuron-cc.rst
│ │ ├── neuron1/
│ │ │ ├── _legacy-labels.rst
│ │ │ ├── neuronrelease/
│ │ │ │ └── previous-content.rst
│ │ │ └── prev/
│ │ │ ├── content.rst
│ │ │ └── rn.rst
│ │ ├── tensorboard-neuron.rst
│ │ ├── tensorflow/
│ │ │ ├── tensorflow-modelserver-neuron/
│ │ │ │ ├── tensorflow-modelserver-neuron-v2.rst
│ │ │ │ ├── tensorflow-modelserver-neuron.rst
│ │ │ │ └── tensorflow-modelserver-neuronx.rst
│ │ │ ├── tensorflow-neuron/
│ │ │ │ ├── tensorflow-neuron-v2.rst
│ │ │ │ └── tensorflow-neuron.rst
│ │ │ └── tensorflow-neuronx/
│ │ │ └── tensorflow-neuronx.rst
│ │ └── torch-neuron.rst
│ ├── components/
│ │ ├── compiler.rst
│ │ ├── containers.rst
│ │ ├── dev-tools.rst
│ │ ├── dlamis.rst
│ │ ├── index.rst
│ │ ├── jax.rst
│ │ ├── nki-lib.rst
│ │ ├── nki.rst
│ │ ├── nxd-core.rst
│ │ ├── nxd-inference.rst
│ │ ├── nxd-training.rst
│ │ ├── pytorch.rst
│ │ └── runtime.rst
│ ├── documentation/
│ │ └── neuron-documentation.rst
│ ├── index.rst
│ ├── prev/
│ │ ├── 2.25.0/
│ │ │ ├── compiler.rst
│ │ │ ├── containers.rst
│ │ │ ├── dlami.rst
│ │ │ ├── docs-and-samples.rst
│ │ │ ├── index.rst
│ │ │ ├── nx-jax.rst
│ │ │ ├── nx-pytorch.rst
│ │ │ ├── nxd-core.rst
│ │ │ ├── nxd-inference.rst
│ │ │ ├── nxd-training.rst
│ │ │ ├── runtime.rst
│ │ │ └── tools.rst
│ │ ├── 2.26.0/
│ │ │ ├── containers.rst
│ │ │ ├── dlami.rst
│ │ │ ├── index.rst
│ │ │ ├── nki.rst
│ │ │ ├── nx-jax.rst
│ │ │ ├── nx-pytorch.rst
│ │ │ ├── nxd-core.rst
│ │ │ ├── nxd-inference.rst
│ │ │ ├── runtime.rst
│ │ │ └── tools.rst
│ │ ├── 2.26.1.rst
│ │ ├── 2.27.0/
│ │ │ ├── compiler.rst
│ │ │ ├── containers.rst
│ │ │ ├── dlami.rst
│ │ │ ├── index.rst
│ │ │ ├── nki-lib.rst
│ │ │ ├── nki.rst
│ │ │ ├── nx-pytorch.rst
│ │ │ ├── nxd-inference.rst
│ │ │ ├── runtime.rst
│ │ │ └── tools.rst
│ │ ├── 2.27.1.rst
│ │ ├── 2.28.0.rst
│ │ ├── 2.28.1.rst
│ │ ├── content.rst
│ │ └── rn.rst
│ └── releasecontent.rst
├── requirements-python310.txt
├── requirements-python38.txt
├── requirements.txt
├── setup/
│ ├── index.rst
│ ├── index.txt-back
│ ├── install-templates/
│ │ ├── al2-python.rst
│ │ ├── inf1/
│ │ │ ├── compile_mode.rst
│ │ │ ├── deploy_mode.rst
│ │ │ ├── develop_mode.rst
│ │ │ ├── dlami-enable-neuron-mxnet.rst
│ │ │ ├── dlami-enable-neuron-pytorch.rst
│ │ │ ├── launch-inf1-ami.rst
│ │ │ ├── launch-inf1-dlami-aws-cli.rst
│ │ │ ├── launch-inf1-dlami.rst
│ │ │ ├── neuron-pip-install.rst
│ │ │ ├── neuron-pip-setup.rst
│ │ │ ├── note-setup-cntr.rst
│ │ │ ├── note-setup-general.rst
│ │ │ ├── note-setup-libnrt-warning.rst
│ │ │ └── tensorboard-plugin-neuron-pip-install.rst
│ │ ├── inf2/
│ │ │ ├── dlami-enable-neuron-pytorch.rst
│ │ │ ├── launch-inf2-dlami.rst
│ │ │ └── note-setup-libnrt-warning.rst
│ │ ├── launch-instance.txt
│ │ ├── launch-trn1-dlami.rst
│ │ ├── trn1/
│ │ │ └── dlami-notes.rst
│ │ └── trn1-ga-warning.txt
│ ├── jax/
│ │ ├── dlami.rst
│ │ ├── dlc.rst
│ │ ├── index.rst
│ │ └── manual.rst
│ ├── jax-neuronx.rst
│ ├── legacy-inf1/
│ │ ├── index.rst
│ │ └── pytorch.rst
│ ├── multiframework-dlami.rst
│ ├── mxnet-neuron.rst
│ ├── notebook/
│ │ ├── running-jupyter-notebook-as-script.rst
│ │ └── setup-jupyter-notebook-steps-troubleshooting.rst
│ ├── pytorch/
│ │ ├── dlami.rst
│ │ ├── dlc.rst
│ │ ├── index.rst
│ │ ├── manual.rst
│ │ ├── update-dlami.rst
│ │ ├── update-dlc.rst
│ │ └── update-manual.rst
│ ├── setup-rocky-linux-9.rst
│ ├── setup-troubleshooting.rst
│ ├── torch-neuron-ubuntu20.rst
│ ├── torch-neuron.rst
│ ├── torch-neuronx.rst
│ └── troubleshooting.rst
├── src/
│ ├── benchmark/
│ │ ├── helper_scripts/
│ │ │ ├── llmperf_dp.patch
│ │ │ ├── llmperf_reasoning.patch
│ │ │ └── neuron_perf.patch
│ │ └── tensorflow/
│ │ ├── distilbert-base-uncased-finetuned-sst-2-english_benchmark.py
│ │ └── distilbert-base-uncased-finetuned-sst-2-english_compile.py
│ ├── examples/
│ │ ├── mxnet/
│ │ │ ├── README.md
│ │ │ ├── data_parallel/
│ │ │ │ ├── benchmark_utils.py
│ │ │ │ ├── data_parallel_tutorial.ipynb
│ │ │ │ └── parallel.py
│ │ │ ├── mxnet-gluon-tutorial.ipynb
│ │ │ ├── resnet50/
│ │ │ │ └── resnet50.ipynb
│ │ │ └── resnet50_neuroncore_groups.ipynb
│ │ ├── neuron-monitor/
│ │ │ └── neuron-monitor-grafana.json
│ │ ├── pytorch/
│ │ │ ├── bert_tutorial/
│ │ │ │ ├── README.md
│ │ │ │ ├── THIRD
│ │ │ │ ├── THIRD PARTY LICENSE.txt
│ │ │ │ ├── bert_benchmark_utils.py
│ │ │ │ ├── glue_mrpc_dev.tsv
│ │ │ │ ├── parallel.py
│ │ │ │ ├── tutorial_pretrained_bert.ipynb
│ │ │ │ └── tutorial_pretrained_bert_shared_weights.ipynb
│ │ │ ├── byoc_sm_bert_tutorial/
│ │ │ │ ├── code/
│ │ │ │ │ └── inference.py
│ │ │ │ ├── container/
│ │ │ │ │ └── Dockerfile
│ │ │ │ └── sagemaker_container_neuron.ipynb
│ │ │ ├── libtorch_demo/
│ │ │ │ ├── bert_neuronx/
│ │ │ │ │ ├── compile.py
│ │ │ │ │ └── detect_instance.py
│ │ │ │ ├── clean.sh
│ │ │ │ ├── example_app/
│ │ │ │ │ ├── README.txt
│ │ │ │ │ ├── build.sh
│ │ │ │ │ ├── core_count.hpp
│ │ │ │ │ ├── example_app.cpp
│ │ │ │ │ ├── utils.cpp
│ │ │ │ │ └── utils.hpp
│ │ │ │ ├── neuron.patch
│ │ │ │ ├── run_tests.sh
│ │ │ │ ├── setup.sh
│ │ │ │ ├── tokenizers_binding/
│ │ │ │ │ ├── build.sh
│ │ │ │ │ ├── remote_rust_tokenizer.h
│ │ │ │ │ ├── run.sh
│ │ │ │ │ ├── run_python.sh
│ │ │ │ │ ├── tokenizer_test
│ │ │ │ │ ├── tokenizer_test.cpp
│ │ │ │ │ └── tokenizer_test.py
│ │ │ │ └── trace_bert_neuron.py
│ │ │ ├── mnist_mlp/
│ │ │ │ ├── train_monitor.py
│ │ │ │ └── train_tb.py
│ │ │ ├── neuronx_distributed/
│ │ │ │ └── t5-inference/
│ │ │ │ ├── t5-inference-tutorial.ipynb
│ │ │ │ ├── t5_model_layers.py
│ │ │ │ ├── t5_models.py
│ │ │ │ └── wrapper.py
│ │ │ ├── pipeline_tutorial/
│ │ │ │ └── neuroncore_pipeline_pytorch.ipynb
│ │ │ ├── resnet50.ipynb
│ │ │ ├── resnet50_partition.ipynb
│ │ │ ├── torch-neuronx/
│ │ │ │ ├── bert-base-cased-finetuned-mrpc-inference-on-trn1-tutorial.ipynb
│ │ │ │ ├── resnet50-inference-on-trn1-tutorial.ipynb
│ │ │ │ └── t5-inference-tutorial.ipynb
│ │ │ ├── torchserve/
│ │ │ │ ├── benchmark_bert.py
│ │ │ │ ├── config.json
│ │ │ │ ├── handler_bert.py
│ │ │ │ ├── handler_bert_neuronx.py
│ │ │ │ ├── infer_bert.py
│ │ │ │ ├── torchserve.config
│ │ │ │ ├── trace_bert_neuron.py
│ │ │ │ └── trace_bert_neuronx.py
│ │ │ ├── transformers-marianmt.ipynb
│ │ │ └── yolo_v4.ipynb
│ │ └── tensorflow/
│ │ ├── bert_demo/
│ │ │ ├── LICENSE
│ │ │ ├── README.md
│ │ │ ├── bert_client.py
│ │ │ ├── bert_model.py
│ │ │ ├── bert_model_server.py
│ │ │ ├── bert_no_model.py
│ │ │ ├── bert_server.py
│ │ │ ├── download_mrpc_data.py
│ │ │ ├── glue_mrpc_dev.tsv
│ │ │ ├── latency_printer.py
│ │ │ ├── mrpc.proto
│ │ │ ├── mrpc_feature.py
│ │ │ ├── mrpc_pb2.py
│ │ │ ├── mrpc_pb2_grpc.py
│ │ │ ├── protoc.sh
│ │ │ ├── setup.py
│ │ │ ├── tokenization.py
│ │ │ ├── tune_save.sh
│ │ │ └── uncased_L-24_H-1024_A-16.vocab.txt
│ │ ├── huggingface_bert/
│ │ │ └── huggingface_bert.ipynb
│ │ ├── k8s_bert_demo/
│ │ │ ├── Dockerfile.tfserving_example
│ │ │ ├── README.md
│ │ │ ├── bert_client.py
│ │ │ └── bert_service.yml
│ │ ├── keras_resnet50/
│ │ │ ├── LICENSE
│ │ │ ├── README.md
│ │ │ ├── fp32tofp16.py
│ │ │ ├── full_sweep
│ │ │ ├── gen_resnet50_keras.py
│ │ │ ├── infer_resnet50_keras.py
│ │ │ ├── infer_resnet50_keras_loadtest.py
│ │ │ ├── keras_resnet50.ipynb
│ │ │ ├── optimize_for_inference.py
│ │ │ ├── pb2sm_compile.py
│ │ │ └── run_all
│ │ ├── openpose_demo/
│ │ │ └── openpose.ipynb
│ │ ├── ssd300_demo/
│ │ │ ├── README.md
│ │ │ ├── ssd300_detection.py
│ │ │ ├── ssd300_evaluation.py
│ │ │ ├── ssd300_evaluation_client.py
│ │ │ └── ssd300_model.py
│ │ ├── tensorflow-neuronx/
│ │ │ └── tfneuronx-roberta-base-tutorial.ipynb
│ │ ├── tensorflow_resnet50/
│ │ │ └── resnet50.ipynb
│ │ ├── tensorflow_serving_tutorial.rst
│ │ ├── yolo_v3_demo/
│ │ │ ├── yolo_v3.ipynb
│ │ │ └── yolo_v3_coco_saved_model.py
│ │ └── yolo_v4_demo/
│ │ ├── README.md
│ │ ├── evaluate.ipynb
│ │ └── yolo_v4_coco_saved_model.py
│ ├── helperscripts/
│ │ ├── installationScripts/
│ │ │ └── python_instructions.txt
│ │ ├── n2-helper.py
│ │ ├── n2-manifest.json
│ │ ├── neuron-releases-manifest.json
│ │ ├── neuron-setup-example.py
│ │ ├── neuronsetuphelper.py
│ │ └── release-manifest-def.py
│ ├── k8/
│ │ ├── bert_service.yml
│ │ ├── k8s-neuron-device-plugin-rbac.yml
│ │ ├── k8s-neuron-device-plugin.yml
│ │ ├── k8s-neuron-monitor-daemonset.yml
│ │ ├── k8s-neuron-scheduler-configmap.yml
│ │ ├── k8s-neuron-scheduler-eks.yml
│ │ ├── k8s-neuron-scheduler.yml
│ │ ├── k8s-ultraserver-init-script.sh
│ │ ├── my-scheduler.yml
│ │ └── neuron-problem-detector/
│ │ ├── k8s-neuron-problem-detector-and-recovery-config.yml
│ │ ├── k8s-neuron-problem-detector-and-recovery-rbac.yml
│ │ └── k8s-neuron-problem-detector-and-recovery.yml
│ ├── libnrt/
│ │ ├── README.md
│ │ └── include/
│ │ ├── ndl/
│ │ │ ├── ndl.h
│ │ │ ├── neuron_driver_shared.h
│ │ │ └── neuron_driver_shared_tensor_batch_op.h
│ │ └── nrt/
│ │ ├── ndebug_stream.h
│ │ ├── nds/
│ │ │ └── neuron_ds.h
│ │ ├── nec.h
│ │ ├── nrt.h
│ │ ├── nrt_async.h
│ │ ├── nrt_async_sendrecv.h
│ │ ├── nrt_experimental.h
│ │ ├── nrt_profile.h
│ │ ├── nrt_status.h
│ │ ├── nrt_sys_trace.h
│ │ └── nrt_version.h
│ ├── neuron-gatherinfo/
│ │ ├── LICENSE
│ │ ├── clear_params_tfpb.py
│ │ ├── mx_neuron_check_model.py
│ │ ├── neuron-gatherinfo.py
│ │ └── tf_neuron_check_model.py
│ └── neuronperf/
│ ├── LICENSE
│ ├── README.md
│ ├── build.sh
│ ├── conf.py
│ ├── model_neuron_b1.csv
│ ├── pyproject.toml
│ ├── src/
│ │ └── neuronperf/
│ │ ├── __init__.py
│ │ ├── __version__.py
│ │ ├── benchmarking.py
│ │ ├── compile_constants.py
│ │ ├── cpu/
│ │ │ ├── __init__.py
│ │ │ └── cpu.py
│ │ ├── logging.py
│ │ ├── model_index.py
│ │ ├── mxnet/
│ │ │ ├── __init__.py
│ │ │ └── mxnet.py
│ │ ├── py.typed
│ │ ├── reporting.py
│ │ ├── scripts/
│ │ │ ├── __init__.py
│ │ │ └── run_benchmark_file.py
│ │ ├── tensorflow/
│ │ │ ├── __init__.py
│ │ │ └── tensorflow.py
│ │ ├── timing.py
│ │ └── torch/
│ │ ├── __init__.py
│ │ └── torch.py
│ └── test/
│ └── test_neuronperf.py
├── static/
│ ├── google673a8c4fbaa024d8.html
│ ├── robots.txt
│ └── sitemap1.xml
└── tools/
├── index.rst
├── neuron-explorer/
│ ├── get-started.rst
│ ├── how-to-link-view-source-code.rst
│ ├── how-to-profile-workload.rst
│ ├── index.rst
│ ├── migration-faq.rst
│ ├── overview-ai-recommendations.rst
│ ├── overview-database-viewer.rst
│ ├── overview-device-profiles.rst
│ ├── overview-hierarchy-view.rst
│ ├── overview-memory-viewer.rst
│ ├── overview-summary-page.rst
│ ├── overview-system-profiles.rst
│ ├── overview-tensor-viewer.rst
│ └── view-perfetto.rst
├── neuron-sys-tools/
│ ├── index.rst
│ ├── nccom-test.rst
│ ├── neuron-ls.rst
│ ├── neuron-monitor-user-guide.rst
│ ├── neuron-sysfs-user-guide.rst
│ └── neuron-top-user-guide.rst
├── profiler/
│ ├── neuron-profile-user-guide.rst
│ └── neuron-profiler-2-0-beta-user-guide.rst
├── tensorboard/
│ ├── getting-started-tensorboard-neuronx-plugin.rst
│ └── index.rst
├── third-party-solutions.rst
└── tutorials/
├── index.rst
├── performance-profiling-vllm.rst
├── torch-neuronx-profiling-with-tb.rst
├── tutorial-neuron-monitor-mnist.rst
└── tutorial-tensorboard-scalars-mnist.rst
SYMBOL INDEX (1534 symbols across 207 files)
FILE: _ext/archive.py
function archive_handler (line 6) | def archive_handler(app):
function setup (line 25) | def setup(app):
FILE: _ext/df_tables.py
class DFTable (line 5) | class DFTable(CSVTable):
method __init__ (line 9) | def __init__(self, name, arguments, options, content, lineno,
method get_csv_data (line 15) | def get_csv_data(self):
method run (line 18) | def run(self):
function setup (line 42) | def setup(app):
FILE: _ext/local_documenter.py
class LocalModuleDocumenter (line 7) | class LocalModuleDocumenter(ModuleDocumenter):
method import_object (line 19) | def import_object(self, *args):
method get_module_members (line 27) | def get_module_members(self):
class LocalFunctionDocumenter (line 38) | class LocalFunctionDocumenter(FunctionDocumenter):
method format_name (line 39) | def format_name(self) -> str:
function setup (line 47) | def setup(app):
FILE: _ext/neuron_tag.py
function _in_list (line 285) | def _in_list(cur_file, file_list):
function _splitall (line 290) | def _splitall(path):
function _get_explicit_override (line 307) | def _get_explicit_override(cur_file):
function _get_page_override (line 453) | def _get_page_override(cur_file):
class NeuronTag (line 586) | class NeuronTag(SphinxDirective):
method run (line 588) | def run(self):
method _render (line 682) | def _render(self, text):
function setup (line 692) | def setup(app):
FILE: _ext/sphinx_plotly_directive.py
function save_plotly_figure (line 182) | def save_plotly_figure(fig, path):
function assign_last_line_into_variable (line 207) | def assign_last_line_into_variable(code, variable_name):
function create_directive_block (line 236) | def create_directive_block(name, arguments, options, content):
function create_code_block (line 290) | def create_code_block(code, language=None):
function strip_last_line (line 301) | def strip_last_line(code):
function ends_with_show (line 324) | def ends_with_show(code):
function _option_boolean (line 357) | def _option_boolean(arg):
function _option_context (line 369) | def _option_context(arg):
function _option_format (line 375) | def _option_format(arg):
function _option_fig_vars (line 379) | def _option_fig_vars(arg):
function mark_plot_labels (line 383) | def mark_plot_labels(app, document):
class PlotlyDirective (line 416) | class PlotlyDirective(Directive):
method run (line 441) | def run(self):
function setup (line 456) | def setup(app):
function contains_doctest (line 489) | def contains_doctest(text):
function unescape_doctest (line 501) | def unescape_doctest(text):
function split_code_at_show (line 521) | def split_code_at_show(text):
class FigureFile (line 625) | class FigureFile:
method __init__ (line 626) | def __init__(self, basename, dirname):
method filename (line 631) | def filename(self, format):
method filenames (line 634) | def filenames(self):
function out_of_date (line 638) | def out_of_date(original, derived):
class PlotError (line 648) | class PlotError(RuntimeError):
function run_code (line 652) | def run_code(code, code_path, ns=None, function_name=None, fig_vars=None):
function get_plot_formats (line 727) | def get_plot_formats(config):
function render_figures (line 745) | def render_figures(
function run (line 855) | def run(arguments, content, options, state_machine, state, lineno):
FILE: _ext/symlink.py
function remove_symlink_handler (line 6) | def remove_symlink_handler(app, exception):
function setup (line 21) | def setup(app):
FILE: _utilities/add_meta.py
function infer_meta (line 12) | def infer_meta(filepath: str, content: str) -> dict:
function has_meta_field (line 67) | def has_meta_field(content: str, field: str) -> bool:
function process_file (line 72) | def process_file(filepath: str, dry_run: bool = False):
function main (line 162) | def main():
FILE: _utilities/audit_frameworks.py
function _resolve_path (line 30) | def _resolve_path(ref: str, referencing_file: Path, root: Path) -> str |...
function _resolve_to_files (line 49) | def _resolve_to_files(base: str, root: Path) -> list[str]:
function extract_toctree_entries (line 71) | def extract_toctree_entries(content: str, filepath: Path, root: Path) ->...
function extract_doc_refs (line 114) | def extract_doc_refs(content: str, filepath: Path, root: Path) -> set[str]:
function extract_include_refs (line 126) | def extract_include_refs(content: str, filepath: Path, root: Path) -> se...
function extract_ref_labels (line 138) | def extract_ref_labels(content: str) -> set[str]:
function extract_label_definitions (line 143) | def extract_label_definitions(content: str) -> set[str]:
function find_all_framework_files (line 152) | def find_all_framework_files(root: Path) -> tuple[set[str], set[str], se...
function collect_all_references (line 178) | def collect_all_references(root: Path) -> tuple[set[str], set[str], set[...
function build_label_to_file_map (line 213) | def build_label_to_file_map(root: Path) -> dict[str, str]:
function detect_orphans (line 234) | def detect_orphans(root: Path) -> list[dict]:
function _check_stale_python (line 300) | def _check_stale_python(content: str) -> list[str]:
function _check_stale_sdk (line 312) | def _check_stale_sdk(content: str) -> list[str]:
function _check_stale_os (line 322) | def _check_stale_os(content: str) -> list[str]:
function _check_torch_neuron_unsupported_os (line 327) | def _check_torch_neuron_unsupported_os(content: str) -> list[str]:
function _check_neuron_cc (line 339) | def _check_neuron_cc(content: str) -> list[str]:
function detect_stale_pages (line 346) | def detect_stale_pages(root: Path) -> list[dict]:
function generate_report (line 398) | def generate_report(orphans: list[dict], stale: list[dict]) -> str:
function main (line 437) | def main():
FILE: _utilities/create_sitemap.py
function create_sitemap (line 8) | def create_sitemap(root_dir, base_url):
FILE: _utilities/format_build_logs.py
function check_python_version (line 20) | def check_python_version():
function check_pip_installed (line 39) | def check_pip_installed():
function find_repo_root (line 50) | def find_repo_root():
function setup_venv (line 77) | def setup_venv(repo_parent):
function get_venv_python (line 103) | def get_venv_python(venv_path):
function get_venv_pip (line 110) | def get_venv_pip(venv_path):
function install_requirements (line 117) | def install_requirements(repo_root, venv_pip):
function run_sphinx_build (line 137) | def run_sphinx_build(repo_root, venv_path):
function parse_build_log (line 189) | def parse_build_log(log_text):
function categorize_issues (line 346) | def categorize_issues(issues):
function format_markdown (line 371) | def format_markdown(errors, warnings, build_time):
function main (line 449) | def main():
FILE: _utilities/inject_archive_meta.py
function find_title_end (line 29) | def find_title_end(lines):
function inject_meta_and_warning (line 61) | def inject_meta_and_warning(filepath, framework="MXNet"):
function main (line 122) | def main():
FILE: _utilities/migrate_setup_content.py
function find_rst_files (line 58) | def find_rst_files(base_dir: str) -> list[Path]:
function find_references (line 70) | def find_references(content: str, file_path: Path) -> list[dict]:
function apply_fix (line 130) | def apply_fix(file_path: Path, refs: list[dict]) -> bool:
function main (line 156) | def main():
FILE: about-neuron/news-and-blogs/validate_articles.py
function validate_url (line 39) | def validate_url(url):
function validate_date (line 51) | def validate_date(date_str):
function validate_article (line 60) | def validate_article(article, index, section):
function main (line 157) | def main():
FILE: archive/neuronperf/test_simple_pt.py
class Model (line 9) | class Model(torch.nn.Module):
method forward (line 10) | def forward(self, x):
FILE: archive/src/benchmark/pytorch/bert-base-cased_benchmark.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/bert-base-cased_compile.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/bert-base-uncased_benchmark.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/bert-base-uncased_compile.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/distilbert-base-uncased-finetuned-sst-2-english_benchmark.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/distilbert-base-uncased-finetuned-sst-2-english_compile.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/distilbert-base-uncased_benchmark.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/distilbert-base-uncased_compile.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/distilroberta-base_benchmark.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/distilroberta-base_compile.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: archive/src/benchmark/pytorch/hf-google-vit_benchmark.py
function benchmark (line 10) | def benchmark(batch_size):
FILE: archive/src/benchmark/pytorch/hf-openai-clip_benchmark.py
function benchmark (line 10) | def benchmark(model_name, batch_size):
FILE: archive/src/benchmark/pytorch/hf_pretrained_wav2vec2_conformer_relpos_benchmark.py
function benchmark (line 9) | def benchmark():
FILE: archive/src/benchmark/pytorch/hf_pretrained_wav2vec2_conformer_rope_benchmark.py
function benchmark (line 9) | def benchmark():
FILE: archive/src/benchmark/pytorch/inf2_benchmark.py
class GPT2Neuron (line 11) | class GPT2Neuron(torch.nn.Module):
method __init__ (line 12) | def __init__(self, model) -> None:
method forward (line 16) | def forward(self, input_ids, attention_mask):
function benchmark (line 19) | def benchmark(model_name, batch_size, sequence_length):
FILE: archive/src/benchmark/pytorch/opt_benchmark.py
class Wrapper (line 23) | class Wrapper(torch.nn.Module):
method __init__ (line 24) | def __init__(self, filename):
method forward (line 32) | def forward(self, *inputs):
function load_fn (line 36) | def load_fn(filename, **kwargs):
function env_setup_fn (line 40) | def env_setup_fn(*_):
function preprocess_fn (line 43) | def preprocess_fn(inputs):
function postprocess_fn (line 46) | def postprocess_fn(outputs):
function benchmark (line 49) | def benchmark():
FILE: archive/src/benchmark/pytorch/perceiver-multimodal_benchmark.py
function benchmark (line 24) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 56) | class LatencyCollector:
method __init__ (line 57) | def __init__(self):
method pre_hook (line 61) | def pre_hook(self, *args):
method hook (line 64) | def hook(self, *args):
method percentile (line 67) | def percentile(self, percent):
class MultimodalPerceiverWrapper (line 76) | class MultimodalPerceiverWrapper(nn.Module):
method __init__ (line 77) | def __init__(self, perceiver_model, nchunks, image_chunk_size, audio_c...
method forward (line 84) | def forward(self, inputs: torch.FloatTensor,
function custom_model_forward (line 162) | def custom_model_forward(
function custom_decoder_query (line 191) | def custom_decoder_query(self, inputs, modality_sizes=None, inputs_witho...
class EncoderWrapper (line 253) | class EncoderWrapper(nn.Module):
method __init__ (line 254) | def __init__(self, encoder):
method forward (line 258) | def forward(self, embedding_output, inputs, extended_attention_mask):
class NeuronEncoder (line 262) | class NeuronEncoder(nn.Module):
method __init__ (line 263) | def __init__(self, encoder_wrapper):
method forward (line 267) | def forward(self,
class DecoderWrapper (line 282) | class DecoderWrapper(nn.Module):
method __init__ (line 283) | def __init__(self, decoder, decoder_query_audio, decoder_query_image, ...
method forward (line 292) | def forward(self, z, query_mask,
class NeuronDecoder (line 318) | class NeuronDecoder(nn.Module):
method __init__ (line 319) | def __init__(self, decoder_wrapper):
method forward (line 325) | def forward(self, z, query_mask, inputs, modality_sizes, inputs_withou...
function autoencode_video (line 382) | def autoencode_video(images, audio, nchunks, image_chunk_size, audio_chu...
FILE: archive/src/benchmark/pytorch/perceiver-multimodal_compile.py
class MultimodalPerceiverWrapper (line 20) | class MultimodalPerceiverWrapper(nn.Module):
method __init__ (line 21) | def __init__(self, perceiver_model, nchunks, image_chunk_size, audio_c...
method forward (line 28) | def forward(self, inputs: torch.FloatTensor,
function custom_model_forward (line 106) | def custom_model_forward(
function custom_decoder_query (line 135) | def custom_decoder_query(self, inputs, modality_sizes=None, inputs_witho...
class EncoderWrapper (line 197) | class EncoderWrapper(nn.Module):
method __init__ (line 198) | def __init__(self, encoder):
method forward (line 202) | def forward(self, embedding_output, inputs, extended_attention_mask):
class NeuronEncoder (line 206) | class NeuronEncoder(nn.Module):
method __init__ (line 207) | def __init__(self, encoder_wrapper):
method forward (line 211) | def forward(self,
class DecoderWrapper (line 226) | class DecoderWrapper(nn.Module):
method __init__ (line 227) | def __init__(self, decoder, decoder_query_audio, decoder_query_image, ...
method forward (line 236) | def forward(self, z, query_mask,
class NeuronDecoder (line 262) | class NeuronDecoder(nn.Module):
method __init__ (line 263) | def __init__(self, decoder_wrapper):
method forward (line 269) | def forward(self, z, query_mask, inputs, modality_sizes, inputs_withou...
FILE: archive/src/benchmark/pytorch/perceiver-vision_benchmark.py
function get_batch (line 16) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/perceiver-vision_compile.py
function get_batch (line 16) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/pixart_alpha_benchmark.py
function benchmark (line 35) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 69) | class LatencyCollector:
method __init__ (line 70) | def __init__(self):
method pre_hook (line 74) | def pre_hook(self, *args):
method hook (line 77) | def hook(self, *args):
method percentile (line 80) | def percentile(self, percent):
class InferenceTextEncoderWrapper (line 89) | class InferenceTextEncoderWrapper(nn.Module):
method __init__ (line 90) | def __init__(self, dtype, t: T5EncoderModel, seqlen: int):
method forward (line 95) | def forward(self, text_input_ids, attention_mask=None):
class InferenceTransformerWrapper (line 98) | class InferenceTransformerWrapper(nn.Module):
method __init__ (line 99) | def __init__(self, transformer: Transformer2DModel):
method forward (line 105) | def forward(self, hidden_states, encoder_hidden_states=None, timestep=...
class SimpleWrapper (line 115) | class SimpleWrapper(nn.Module):
method __init__ (line 116) | def __init__(self, model):
method forward (line 119) | def forward(self, x):
function get_pipe (line 124) | def get_pipe(resolution, dtype):
FILE: archive/src/benchmark/pytorch/pixart_sigma_benchmark.py
function benchmark (line 34) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 68) | class LatencyCollector:
method __init__ (line 69) | def __init__(self):
method pre_hook (line 73) | def pre_hook(self, *args):
method hook (line 76) | def hook(self, *args):
method percentile (line 79) | def percentile(self, percent):
class InferenceTextEncoderWrapper (line 88) | class InferenceTextEncoderWrapper(nn.Module):
method __init__ (line 89) | def __init__(self, dtype, t: T5EncoderModel, seqlen: int):
method forward (line 94) | def forward(self, text_input_ids, attention_mask=None):
class InferenceTransformerWrapper (line 97) | class InferenceTransformerWrapper(nn.Module):
method __init__ (line 98) | def __init__(self, transformer: Transformer2DModel):
method forward (line 104) | def forward(self, hidden_states, encoder_hidden_states=None, timestep=...
class SimpleWrapper (line 114) | class SimpleWrapper(nn.Module):
method __init__ (line 115) | def __init__(self, model):
method forward (line 118) | def forward(self, x):
function get_pipe (line 123) | def get_pipe(resolution, dtype):
FILE: archive/src/benchmark/pytorch/resnet50_benchmark.py
function get_batch (line 12) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/resnet50_compile.py
function get_batch (line 14) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/resnet_benchmark.py
function get_batch (line 12) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/resnet_compile.py
function get_batch (line 12) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/sd2_512_benchmark.py
function benchmark (line 22) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 56) | class LatencyCollector:
method __init__ (line 57) | def __init__(self):
method pre_hook (line 61) | def pre_hook(self, *args):
method hook (line 64) | def hook(self, *args):
method percentile (line 67) | def percentile(self, percent):
class UNetWrap (line 77) | class UNetWrap(nn.Module):
method __init__ (line 78) | def __init__(self, unet):
method forward (line 82) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronUNet (line 86) | class NeuronUNet(nn.Module):
method __init__ (line 87) | def __init__(self, unetwrap):
method forward (line 94) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronTextEncoder (line 98) | class NeuronTextEncoder(nn.Module):
method __init__ (line 99) | def __init__(self, text_encoder):
method forward (line 106) | def forward(self, emb, attention_mask = None):
function decode_latents (line 109) | def decode_latents(self, latents):
class NeuronTypeConversionWrapper (line 135) | class NeuronTypeConversionWrapper(nn.Module):
method __init__ (line 136) | def __init__(self, network):
method forward (line 140) | def forward(self, x):
FILE: archive/src/benchmark/pytorch/sd2_512_compile.py
class UNetWrap (line 26) | class UNetWrap(nn.Module):
method __init__ (line 27) | def __init__(self, unet):
method forward (line 31) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronUNet (line 35) | class NeuronUNet(nn.Module):
method __init__ (line 36) | def __init__(self, unetwrap):
method forward (line 43) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronTextEncoder (line 47) | class NeuronTextEncoder(nn.Module):
method __init__ (line 48) | def __init__(self, text_encoder):
method forward (line 55) | def forward(self, emb, attention_mask = None):
function get_attention_scores (line 60) | def get_attention_scores(self, query, key, attn_mask):
function custom_badbmm (line 95) | def custom_badbmm(a, b):
FILE: archive/src/benchmark/pytorch/sd2_768_benchmark.py
function benchmark (line 22) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 56) | class LatencyCollector:
method __init__ (line 57) | def __init__(self):
method pre_hook (line 61) | def pre_hook(self, *args):
method hook (line 64) | def hook(self, *args):
method percentile (line 67) | def percentile(self, percent):
class UNetWrap (line 77) | class UNetWrap(nn.Module):
method __init__ (line 78) | def __init__(self, unet):
method forward (line 82) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronUNet (line 86) | class NeuronUNet(nn.Module):
method __init__ (line 87) | def __init__(self, unetwrap):
method forward (line 94) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronTextEncoder (line 98) | class NeuronTextEncoder(nn.Module):
method __init__ (line 99) | def __init__(self, text_encoder):
method forward (line 106) | def forward(self, emb, attention_mask = None):
FILE: archive/src/benchmark/pytorch/sd2_768_compile.py
class UNetWrap (line 24) | class UNetWrap(nn.Module):
method __init__ (line 25) | def __init__(self, unet):
method forward (line 29) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronUNet (line 33) | class NeuronUNet(nn.Module):
method __init__ (line 34) | def __init__(self, unetwrap):
method forward (line 41) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronTextEncoder (line 45) | class NeuronTextEncoder(nn.Module):
method __init__ (line 46) | def __init__(self, text_encoder):
method forward (line 53) | def forward(self, emb, attention_mask = None):
function get_attention_scores (line 58) | def get_attention_scores(self, query, key, attn_mask):
function custom_badbmm (line 93) | def custom_badbmm(a, b):
FILE: archive/src/benchmark/pytorch/sd2_inpainting_benchmark.py
function parse_argsuments (line 15) | def parse_argsuments():
class UNetWrap (line 24) | class UNetWrap(nn.Module):
method __init__ (line 25) | def __init__(self, unet):
method forward (line 29) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronUNet (line 33) | class NeuronUNet(nn.Module):
method __init__ (line 34) | def __init__(self, unetwrap):
method forward (line 41) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronTextEncoder (line 45) | class NeuronTextEncoder(nn.Module):
method __init__ (line 46) | def __init__(self, text_encoder):
method forward (line 53) | def forward(self, emb, attention_mask = None):
function get_attention_scores (line 57) | def get_attention_scores(self, query, key, attn_mask):
function custom_badbmm (line 91) | def custom_badbmm(a, b):
function trace_vae_encoder (line 101) | def trace_vae_encoder(model_id, height, width):
function trace_unet (line 122) | def trace_unet(model_id, height, width):
function main (line 158) | def main():
FILE: archive/src/benchmark/pytorch/sd2_inpainting_inference.py
function parse_argsuments (line 20) | def parse_argsuments():
function benchmark (line 32) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 66) | class LatencyCollector:
method __init__ (line 67) | def __init__(self):
method pre_hook (line 71) | def pre_hook(self, *args):
method hook (line 74) | def hook(self, *args):
method percentile (line 77) | def percentile(self, percent):
class UNetWrap (line 90) | class UNetWrap(nn.Module):
method __init__ (line 91) | def __init__(self, unet):
method forward (line 95) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronUNet (line 99) | class NeuronUNet(nn.Module):
method __init__ (line 100) | def __init__(self, unetwrap):
method forward (line 107) | def forward(self, sample, timestep, encoder_hidden_states, timestep_co...
class NeuronTextEncoder (line 111) | class NeuronTextEncoder(nn.Module):
method __init__ (line 112) | def __init__(self, text_encoder):
method forward (line 119) | def forward(self, emb, attention_mask = None):
function get_attention_scores (line 123) | def get_attention_scores(self, query, key, attn_mask):
function custom_badbmm (line 157) | def custom_badbmm(a, b):
function main (line 163) | def main():
FILE: archive/src/benchmark/pytorch/sd_15_512_benchmark.py
function benchmark (line 21) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 55) | class LatencyCollector:
method __init__ (line 56) | def __init__(self):
method pre_hook (line 60) | def pre_hook(self, *args):
method hook (line 63) | def hook(self, *args):
method percentile (line 66) | def percentile(self, percent):
class UNetWrap (line 76) | class UNetWrap(nn.Module):
method __init__ (line 77) | def __init__(self, unet):
method forward (line 81) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronUNet (line 85) | class NeuronUNet(nn.Module):
method __init__ (line 86) | def __init__(self, unetwrap):
method forward (line 93) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronTextEncoder (line 97) | class NeuronTextEncoder(nn.Module):
method __init__ (line 98) | def __init__(self, text_encoder):
method forward (line 105) | def forward(self, emb, attention_mask = None):
class NeuronSafetyModelWrap (line 109) | class NeuronSafetyModelWrap(nn.Module):
method __init__ (line 110) | def __init__(self, safety_model):
method forward (line 114) | def forward(self, clip_inputs):
FILE: archive/src/benchmark/pytorch/sd_15_512_compile.py
function get_attention_scores (line 24) | def get_attention_scores(self, query, key, attn_mask):
function cust_badbmm (line 59) | def cust_badbmm(a, b, scale):
class UNetWrap (line 65) | class UNetWrap(nn.Module):
method __init__ (line 66) | def __init__(self, unet):
method forward (line 70) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronUNet (line 74) | class NeuronUNet(nn.Module):
method __init__ (line 75) | def __init__(self, unetwrap):
method forward (line 82) | def forward(self, sample, timestep, encoder_hidden_states, cross_atten...
class NeuronTextEncoder (line 86) | class NeuronTextEncoder(nn.Module):
method __init__ (line 87) | def __init__(self, text_encoder):
method forward (line 94) | def forward(self, emb, attention_mask = None):
class NeuronSafetyModelWrap (line 98) | class NeuronSafetyModelWrap(nn.Module):
method __init__ (line 99) | def __init__(self, safety_model):
method forward (line 103) | def forward(self, clip_inputs):
FILE: archive/src/benchmark/pytorch/sd_4x_upscaler_benchmark.py
class UNetWrap (line 21) | class UNetWrap(nn.Module):
method __init__ (line 22) | def __init__(self, unet):
method forward (line 26) | def forward(
class NeuronUNet (line 40) | class NeuronUNet(nn.Module):
method __init__ (line 41) | def __init__(self, unetwrap):
method forward (line 48) | def forward(
class NeuronTextEncoder (line 66) | class NeuronTextEncoder(nn.Module):
method __init__ (line 67) | def __init__(self, text_encoder):
method forward (line 74) | def forward(self, emb, attention_mask=None):
function benchmark (line 82) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 116) | class LatencyCollector:
method __init__ (line 117) | def __init__(self):
method pre_hook (line 121) | def pre_hook(self, *args):
method hook (line 124) | def hook(self, *args):
method percentile (line 127) | def percentile(self, percent):
FILE: archive/src/benchmark/pytorch/sd_4x_upscaler_compile.py
function apply_neuron_attn_override (line 20) | def apply_neuron_attn_override(
function get_attention_scores_neuron (line 40) | def get_attention_scores_neuron(self, query, key, attn_mask):
function cust_badbmm (line 52) | def cust_badbmm(a, b, scale):
function neuron_scaled_dot_product_attention (line 58) | def neuron_scaled_dot_product_attention(
class UNetWrap (line 92) | class UNetWrap(nn.Module):
method __init__ (line 93) | def __init__(self, unet):
method forward (line 97) | def forward(
class NeuronUNet (line 111) | class NeuronUNet(nn.Module):
method __init__ (line 112) | def __init__(self, unetwrap):
method forward (line 119) | def forward(
class NeuronTextEncoder (line 137) | class NeuronTextEncoder(nn.Module):
method __init__ (line 138) | def __init__(self, text_encoder):
method forward (line 145) | def forward(self, emb, attention_mask=None):
FILE: archive/src/benchmark/pytorch/sdxl_base_1024_benchmark.py
function benchmark (line 22) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 56) | class LatencyCollector:
method __init__ (line 57) | def __init__(self):
method pre_hook (line 61) | def pre_hook(self, *args):
method hook (line 64) | def hook(self, *args):
method percentile (line 67) | def percentile(self, percent):
class UNetWrap (line 76) | class UNetWrap(nn.Module):
method __init__ (line 77) | def __init__(self, unet):
method forward (line 81) | def forward(self, sample, timestep, encoder_hidden_states, text_embeds...
class NeuronUNet (line 90) | class NeuronUNet(nn.Module):
method __init__ (line 91) | def __init__(self, unetwrap):
method forward (line 99) | def forward(self, sample, timestep, encoder_hidden_states, added_cond_...
class TextEncoderOutputWrapper (line 107) | class TextEncoderOutputWrapper(nn.Module):
method __init__ (line 108) | def __init__(self, traceable_text_encoder, original_text_encoder):
method forward (line 115) | def forward(self, text_input_ids, output_hidden_states=True):
method __init__ (line 121) | def __init__(self, traceable_text_encoder, original_text_encoder):
method forward (line 128) | def forward(self, text_input_ids, output_hidden_states=True):
class TextEncoderOutputWrapper (line 120) | class TextEncoderOutputWrapper(nn.Module):
method __init__ (line 108) | def __init__(self, traceable_text_encoder, original_text_encoder):
method forward (line 115) | def forward(self, text_input_ids, output_hidden_states=True):
method __init__ (line 121) | def __init__(self, traceable_text_encoder, original_text_encoder):
method forward (line 128) | def forward(self, text_input_ids, output_hidden_states=True):
class TraceableTextEncoder (line 132) | class TraceableTextEncoder(nn.Module):
method __init__ (line 133) | def __init__(self, text_encoder):
method forward (line 137) | def forward(self, text_input_ids):
FILE: archive/src/benchmark/pytorch/sdxl_base_1024_compile.py
function apply_neuron_attn_override (line 18) | def apply_neuron_attn_override(
function get_attention_scores_neuron (line 41) | def get_attention_scores_neuron(self, query, key, attn_mask):
function custom_badbmm (line 60) | def custom_badbmm(a, b, scale):
function neuron_scaled_dot_product_attention (line 65) | def neuron_scaled_dot_product_attention(
class UNetWrap (line 103) | class UNetWrap(nn.Module):
method __init__ (line 104) | def __init__(self, unet):
method forward (line 108) | def forward(
class NeuronUNet (line 121) | class NeuronUNet(nn.Module):
method __init__ (line 122) | def __init__(self, unetwrap):
method forward (line 130) | def forward(
class TextEncoderOutputWrapper (line 148) | class TextEncoderOutputWrapper(nn.Module):
method __init__ (line 149) | def __init__(self, traceable_text_encoder, original_text_encoder):
method forward (line 156) | def forward(self, text_input_ids, output_hidden_states=True):
class TraceableTextEncoder (line 160) | class TraceableTextEncoder(nn.Module):
method __init__ (line 161) | def __init__(self, text_encoder):
method forward (line 165) | def forward(self, text_input_ids):
FILE: archive/src/benchmark/pytorch/sdxl_base_and_refiner_1024_benchmark.py
function benchmark (line 21) | def benchmark(n_runs, test_name, model, model_inputs):
class LatencyCollector (line 55) | class LatencyCollector:
method __init__ (line 56) | def __init__(self):
method pre_hook (line 60) | def pre_hook(self, *args):
method hook (line 63) | def hook(self, *args):
method percentile (line 66) | def percentile(self, percent):
class UNetWrap (line 75) | class UNetWrap(nn.Module):
method __init__ (line 76) | def __init__(self, unet):
method forward (line 80) | def forward(self, sample, timestep, encoder_hidden_states, text_embeds...
class NeuronUNet (line 89) | class NeuronUNet(nn.Module):
method __init__ (line 90) | def __init__(self, unetwrap):
method forward (line 98) | def forward(self, sample, timestep, encoder_hidden_states, added_cond_...
function run_refiner_and_base (line 107) | def run_refiner_and_base(base, refiner, prompt, n_steps=40, high_noise_f...
FILE: archive/src/benchmark/pytorch/sdxl_base_and_refiner_1024_compile.py
function get_attention_scores_neuron (line 16) | def get_attention_scores_neuron(self, query, key, attn_mask):
function custom_badbmm (line 36) | def custom_badbmm(a, b, scale):
class UNetWrap (line 42) | class UNetWrap(nn.Module):
method __init__ (line 43) | def __init__(self, unet):
method forward (line 47) | def forward(self, sample, timestep, encoder_hidden_states, text_embeds...
class NeuronUNet (line 56) | class NeuronUNet(nn.Module):
method __init__ (line 57) | def __init__(self, unetwrap):
method forward (line 65) | def forward(self, sample, timestep, encoder_hidden_states, added_cond_...
FILE: archive/src/benchmark/pytorch/unet_benchmark.py
function get_batch (line 11) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/unet_compile.py
function get_batch (line 11) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/vgg_benchmark.py
function get_batch (line 12) | def get_batch(batch_size):
FILE: archive/src/benchmark/pytorch/vgg_compile.py
function get_batch (line 12) | def get_batch(batch_size):
FILE: archive/torch-neuron/placement.py
function set_neuron_cores (line 42) | def set_neuron_cores(trace: 'torch.jit.ScriptModule', start_nc: int=-1, ...
function set_multicore (line 104) | def set_multicore(trace: 'torch.jit.ScriptModule'):
function neuron_cores_context (line 142) | def neuron_cores_context(start_nc: int=-1, nc_count: int=-1):
function multicore_context (line 208) | def multicore_context():
FILE: archive/tutorials/ssd300_demo/ssd300_detection.py
function main (line 13) | def main():
FILE: archive/tutorials/ssd300_demo/ssd300_evaluation.py
function get_val_dataset (line 19) | def get_val_dataset(val_annotate, val_coco_root):
function main (line 26) | def main():
FILE: archive/tutorials/ssd300_demo/ssd300_evaluation_client.py
function get_val_dataset (line 21) | def get_val_dataset(val_annotate, val_coco_root):
function main (line 28) | def main():
FILE: archive/tutorials/ssd300_demo/ssd300_model.py
function decode_jpeg_resize (line 19) | def decode_jpeg_resize(input_tensor, image_size):
function preprocessor (line 35) | def preprocessor(input_tensor, image_size):
function tf_Conv2d (line 43) | def tf_Conv2d(input_tensor, module, first_conv=False):
function tf_BatchNorm2d (line 60) | def tf_BatchNorm2d(input_tensor, module):
function tf_MaxPool2d (line 69) | def tf_MaxPool2d(input_tensor, module):
function tf_Bottleneck (line 74) | def tf_Bottleneck(input_tensor, module):
function tf_SequentialBottleneck (line 88) | def tf_SequentialBottleneck(tensor, seq, resnet):
function tf_bbox_view (line 95) | def tf_bbox_view(detection_feed, modules, ndim):
function tf_feature_extractor (line 111) | def tf_feature_extractor(input_tensor, resnet):
function tf_box_predictor (line 128) | def tf_box_predictor(tensor, ssd300_torch):
function tf_ssd300 (line 148) | def tf_ssd300(input_tensor, ssd300_torch):
function scale_back_batch (line 155) | def scale_back_batch(bboxes_in, scores_in, scale_xy, scale_wh, dboxes_xy...
function select_nms_outputs (line 180) | def select_nms_outputs(input_tensors):
function postprocessor (line 184) | def postprocessor(ploc_ts, plabel_ts, bbox_scale_hw_ts, scale_xy, scale_...
class DefaultBoxes (line 215) | class DefaultBoxes(object):
method __init__ (line 217) | def __init__(self, fig_size, feat_size, steps, scales, aspect_ratios,
method scale_xy (line 262) | def scale_xy(self):
method scale_wh (line 266) | def scale_wh(self):
method __call__ (line 269) | def __call__(self, order="ltrb"):
function dboxes300_coco (line 274) | def dboxes300_coco():
function main (line 285) | def main():
FILE: conf.py
function get_env_vars_from_gh (line 25) | def get_env_vars_from_gh():
function get_env_vars_from_rtd (line 33) | def get_env_vars_from_rtd():
function get_env_vars (line 44) | def get_env_vars():
FILE: containers/docker-example/training/mlp_train.py
function main (line 22) | def main():
FILE: containers/docker-example/training/model.py
class MLP (line 5) | class MLP(nn.Module):
method __init__ (line 6) | def __init__(self, input_size = 28 * 28, output_size = 10, layers = [1...
method forward (line 12) | def forward(self, x):
FILE: nki/_ext/nki_directives.py
class NKIExampleReader (line 37) | class NKIExampleReader(LiteralIncludeReader):
method __init__ (line 39) | def __init__(self, filename: str, options: dict[str, Any], config: Con...
method nki_example_filter (line 49) | def nki_example_filter(
method read (line 86) | def read(self, location: Union[tuple[str, int], None] = None) -> tuple...
class NKIExample (line 106) | class NKIExample(LiteralInclude):
method run (line 116) | def run(self) -> list[Node]:
function setup (line 180) | def setup(app: Sphinx) -> ExtensionMetadata:
FILE: nki/api/nki/__init__.py
function jit (line 6) | def jit(func=None, mode="auto", **kwargs):
function simulate (line 58) | def simulate(kernel):
FILE: nki/api/nki/collectives/__init__.py
class NKIObject (line 6) | class NKIObject:
class ReplicaGroup (line 11) | class ReplicaGroup(NKIObject):
function all_gather (line 18) | def all_gather(srcs, dsts, replica_group, collective_dim):
function all_reduce (line 39) | def all_reduce(srcs, dsts, replica_group, op):
function all_to_all (line 57) | def all_to_all(srcs, dsts, replica_group, collective_dim):
function all_to_all_v (line 74) | def all_to_all_v(srcs, dsts, replica_group, metadata_tensor, recv_counts...
function collective_permute (line 93) | def collective_permute(srcs, dsts, source_target_pairs):
function collective_permute_implicit (line 114) | def collective_permute_implicit(srcs_by_channel, dsts_by_channel, replic...
function collective_permute_implicit_current_processing_rank_id (line 144) | def collective_permute_implicit_current_processing_rank_id(iteration_id,...
function collective_permute_implicit_reduce (line 180) | def collective_permute_implicit_reduce(srcs0_by_channel, srcs1_by_channe...
function rank_id (line 214) | def rank_id():
function reduce_scatter (line 221) | def reduce_scatter(srcs, dsts, replica_group, collective_dim, op):
FILE: nki/api/nki/isa/__init__.py
class NKIObject (line 8) | class NKIObject:
class NkiValidationError (line 13) | class NkiValidationError(Exception):
class VirtualRegister (line 18) | class VirtualRegister(NKIObject):
class dge_mode (line 66) | class dge_mode(Enum):
class dma_engine (line 79) | class dma_engine(Enum):
class engine (line 90) | class engine(Enum):
class matmul_perf_mode (line 109) | class matmul_perf_mode(Enum):
class nc_version (line 118) | class nc_version(Enum):
class oob_mode (line 129) | class oob_mode(Enum):
class reduce_cmd (line 138) | class reduce_cmd(Enum):
function activation (line 153) | def activation(dst, op, data, bias=None, scale=1.0, reduce_op=None, redu...
function activation_reduce (line 248) | def activation_reduce(dst, op, data, reduce_op, reduce_res, bias=None, s...
function affine_select (line 300) | def affine_select(dst, pattern, channel_multiplier, on_true_tile, on_fal...
function bn_aggr (line 375) | def bn_aggr(dst, data, name=None):
function bn_stats (line 398) | def bn_stats(dst, data, name=None):
function core_barrier (line 435) | def core_barrier(data, cores, engine=engine.gpsimd, name=None):
function dma_compute (line 491) | def dma_compute(dst, srcs, reduce_op, scales=None, unique_indices=True, ...
function dma_copy (line 561) | def dma_copy(dst, src, oob_mode=oob_mode.error, dge_mode=dge_mode.unknow...
function dma_transpose (line 697) | def dma_transpose(dst, src, axes=None, dge_mode=dge_mode.unknown, oob_mo...
function dropout (line 779) | def dropout(dst, data, prob, name=None):
function exponential (line 810) | def exponential(dst, src, max_value=0.0, reduce_res=None, reduce_cmd=red...
function get_nc_version (line 890) | def get_nc_version():
function iota (line 899) | def iota(dst, pattern, offset=0, channel_multiplier=0, name=None):
function local_gather (line 956) | def local_gather(dst, src_buffer, index, num_elem_per_idx=1, num_valid_i...
function max8 (line 1010) | def max8(dst, src, name=None):
function memset (line 1030) | def memset(dst, value, engine=engine.unknown, name=None):
function nc_find_index8 (line 1046) | def nc_find_index8(dst, data, vals, name=None):
function nc_match_replace8 (line 1070) | def nc_match_replace8(dst, data, vals, imm, dst_idx=None, name=None):
function nc_matmul (line 1083) | def nc_matmul(dst, stationary, moving, is_stationary_onezero=False, is_m...
function nc_matmul_mx (line 1242) | def nc_matmul_mx(dst, stationary, moving, stationary_scale, moving_scale...
function nc_n_gather (line 1336) | def nc_n_gather(dst, data, indices, name=None):
function nc_stream_shuffle (line 1385) | def nc_stream_shuffle(dst, src, shuffle_mask, name=None):
function nc_transpose (line 1421) | def nc_transpose(dst, data, engine=engine.unknown, name=None):
function nonzero_with_count (line 1462) | def nonzero_with_count(dst, src, index_offset=0, padding_val=-1, name=No...
function quantize_mx (line 1563) | def quantize_mx(dst, src, dst_scale, name=None):
function rand2 (line 1610) | def rand2(dst, min, max, name=None):
function rand_get_state (line 1652) | def rand_get_state(dst, engine=engine.unknown, name=None):
function rand_set_state (line 1683) | def rand_set_state(src_seeds, engine=engine.unknown, name=None):
function range_select (line 1719) | def range_select(dst, on_true_tile, comp_op0, comp_op1, bound0, bound1, ...
function reciprocal (line 1795) | def reciprocal(dst, data, name=None):
function register_alloc (line 1823) | def register_alloc(x=None):
function register_load (line 1866) | def register_load(dst, src):
function register_move (line 1892) | def register_move(dst, src):
function register_store (line 1921) | def register_store(dst, src):
function rng (line 1947) | def rng(dst, engine=engine.unknown, name=None):
function scalar_tensor_tensor (line 1988) | def scalar_tensor_tensor(dst, data, op0, operand0, op1, operand1, revers...
function select_reduce (line 2047) | def select_reduce(dst, predicate, on_true, on_false, reduce_res=None, re...
function sendrecv (line 2112) | def sendrecv(src, dst, send_to_rank, recv_from_rank, pipe_id, dma_engine...
function sequence_bounds (line 2198) | def sequence_bounds(dst, segment_ids, name=None):
function set_rng_seed (line 2222) | def set_rng_seed(src_seeds, name=None):
function tensor_copy (line 2247) | def tensor_copy(dst, src, engine=engine.unknown, name=None):
function tensor_copy_predicated (line 2275) | def tensor_copy_predicated(dst, src, predicate, reverse_pred=False, name...
function tensor_partition_reduce (line 2305) | def tensor_partition_reduce(dst, op, data, name=None):
function tensor_reduce (line 2314) | def tensor_reduce(dst, op, data, axis, negate=False, keepdims=False, nam...
function tensor_scalar (line 2382) | def tensor_scalar(dst, data, op0, operand0, reverse0=False, op1=None, op...
function tensor_scalar_cumulative (line 2439) | def tensor_scalar_cumulative(dst, src, op0, op1, imm0, imm1=None, reduce...
function tensor_scalar_reduce (line 2498) | def tensor_scalar_reduce(dst, data, op0, operand0, reduce_op, reduce_res...
function tensor_tensor (line 2534) | def tensor_tensor(dst, data1, data2, op, engine=engine.unknown, name=None):
function tensor_tensor_scan (line 2578) | def tensor_tensor_scan(dst, data0, data1, initial, op0, op1, reverse0=Fa...
FILE: nki/api/nki/language/__init__.py
class MemoryRegion (line 5) | class MemoryRegion(Enum):
class NKIObject (line 14) | class NKIObject:
class tile_size (line 19) | class tile_size:
class NkiTensor (line 39) | class NkiTensor(NKIObject):
function abs (line 52) | def abs(x, dtype=None):
function add (line 83) | def add(x, y, dtype=None):
function affine_range (line 119) | def affine_range(start, stop=None, step=1):
function all (line 149) | def all(x, axis, dtype=None):
function arctan (line 167) | def arctan(x, dtype=None):
function bitwise_and (line 198) | def bitwise_and(x, y, dtype=None):
function bitwise_or (line 216) | def bitwise_or(x, y, dtype=None):
function bitwise_xor (line 234) | def bitwise_xor(x, y, dtype=None):
function broadcast_to (line 256) | def broadcast_to(x, shape, dtype=None):
function ceil (line 276) | def ceil(x, dtype=None):
function copy (line 311) | def copy(x, dtype=None):
function cos (line 329) | def cos(x, dtype=None):
function device_print (line 356) | def device_print(print_prefix, tensor):
function divide (line 372) | def divide(x, y, dtype=None):
function dropout (line 389) | def dropout(x, rate, dtype=None):
function ds (line 404) | def ds(start, size):
function dynamic_range (line 413) | def dynamic_range(start, stop=None, step=1):
function empty_like (line 438) | def empty_like(x, dtype=None, buffer=None, name=''):
function equal (line 455) | def equal(x, y, dtype=None):
function erf (line 472) | def erf(x, dtype=None):
function erf_dx (line 477) | def erf_dx(x, dtype=None):
function exp (line 482) | def exp(x, dtype=None):
function expand_dims (line 511) | def expand_dims(x, axis):
function floor (line 560) | def floor(x, dtype=None):
function fmod (line 595) | def fmod(x, y, dtype=None):
function full (line 615) | def full(shape, fill_value, dtype, buffer=MemoryRegion.sbuf, name=''):
function gather_flattened (line 631) | def gather_flattened(data, indices, axis=0, dtype=None):
function gelu (line 663) | def gelu(x, dtype=None):
function gelu_apprx_sigmoid (line 668) | def gelu_apprx_sigmoid(x, dtype=None):
function gelu_apprx_sigmoid_dx (line 673) | def gelu_apprx_sigmoid_dx(x, dtype=None):
function gelu_apprx_tanh (line 678) | def gelu_apprx_tanh(x, dtype=None):
function gelu_dx (line 683) | def gelu_dx(x, dtype=None):
function greater (line 688) | def greater(x, y, dtype=None):
function greater_equal (line 705) | def greater_equal(x, y, dtype=None):
function invert (line 737) | def invert(x, dtype=None):
function is_hbm (line 754) | def is_hbm(buffer):
function is_on_chip (line 759) | def is_on_chip(buffer):
function is_psum (line 764) | def is_psum(buffer):
function is_sbuf (line 769) | def is_sbuf(buffer):
function left_shift (line 774) | def left_shift(x, y, dtype=None):
function less (line 792) | def less(x, y, dtype=None):
function less_equal (line 809) | def less_equal(x, y, dtype=None):
function load (line 826) | def load(src, dtype=None):
function load_transpose2d (line 839) | def load_transpose2d(src, dtype=None):
function log (line 852) | def log(x, dtype=None):
function logical_and (line 882) | def logical_and(x, y, dtype=None):
function logical_not (line 900) | def logical_not(x, dtype=None):
function logical_or (line 918) | def logical_or(x, y, dtype=None):
function logical_xor (line 936) | def logical_xor(x, y, dtype=None):
function matmul (line 954) | def matmul(x, y, transpose_x=False):
function max (line 987) | def max(x, axis, dtype=None, keepdims=False):
function maximum (line 1006) | def maximum(x, y, dtype=None):
function mean (line 1041) | def mean(x, axis, dtype=None, keepdims=False):
function min (line 1061) | def min(x, axis, dtype=None, keepdims=False):
function minimum (line 1080) | def minimum(x, y, dtype=None):
function mish (line 1115) | def mish(x, dtype=None):
function mod (line 1120) | def mod(x, y, dtype=None):
function multiply (line 1140) | def multiply(x, y, dtype=None):
function ndarray (line 1175) | def ndarray(shape, dtype, buffer=MemoryRegion.sbuf, name='', address=None):
function negative (line 1187) | def negative(x, dtype=None):
function no_reorder (line 1220) | def no_reorder():
function not_equal (line 1249) | def not_equal(x, y, dtype=None):
function num_programs (line 1266) | def num_programs(axes=0):
function ones (line 1275) | def ones(shape, dtype, buffer=MemoryRegion.sbuf, name=''):
function power (line 1292) | def power(x, y, dtype=None):
function prod (line 1324) | def prod(x, axis, dtype=None, keepdims=False):
function program_id (line 1343) | def program_id(axis=0):
function program_ndim (line 1351) | def program_ndim():
function rand (line 1361) | def rand(shape, dtype, buffer=MemoryRegion.sbuf, name=''):
function random_seed (line 1387) | def random_seed(seed):
function reciprocal (line 1420) | def reciprocal(x, dtype=None):
function relu (line 1451) | def relu(x, dtype=None):
function right_shift (line 1456) | def right_shift(x, y, dtype=None):
function rms_norm (line 1474) | def rms_norm(x, w, axis, n, epsilon=1e-06, dtype=None, compute_dtype=None):
function rsqrt (line 1503) | def rsqrt(x, dtype=None):
function sequential_range (line 1537) | def sequential_range(start, stop=None, step=1):
function shared_identity_matrix (line 1570) | def shared_identity_matrix(n, dtype='uint8', dst=None):
function sigmoid (line 1594) | def sigmoid(x, dtype=None):
function sign (line 1599) | def sign(x, dtype=None):
function silu (line 1630) | def silu(x, dtype=None):
function silu_dx (line 1635) | def silu_dx(x, dtype=None):
function sin (line 1640) | def sin(x, dtype=None):
function softmax (line 1669) | def softmax(x, axis=-1, dtype=None):
function softplus (line 1695) | def softplus(x, dtype=None):
function sqrt (line 1700) | def sqrt(x, dtype=None):
function square (line 1729) | def square(x, dtype=None):
function static_range (line 1758) | def static_range(start, stop=None, step=1):
function store (line 1785) | def store(dst, value):
function subtract (line 1797) | def subtract(x, y, dtype=None):
function sum (line 1832) | def sum(x, axis, dtype=None, keepdims=False):
function tan (line 1851) | def tan(x, dtype=None):
function tanh (line 1880) | def tanh(x, dtype=None):
function transpose (line 1889) | def transpose(x, dtype=None):
function trunc (line 1916) | def trunc(x, dtype=None):
function var (line 1964) | def var(x, axis, dtype=None, keepdims=False):
function where (line 1983) | def where(condition, x, y, dtype=None):
function zeros (line 2030) | def zeros(shape, dtype, buffer=MemoryRegion.sbuf, name=''):
function zeros_like (line 2047) | def zeros_like(x, dtype=None, buffer=None, name=''):
FILE: nki/deep-dives/src/mxfp-matmul/mx_cpu_utils.py
function validate_quantized_dtype (line 9) | def validate_quantized_dtype(dtype):
function get_float32_exp (line 15) | def get_float32_exp(float_data):
function get_mx_fp_max (line 23) | def get_mx_fp_max(mx_dtype):
function get_mx_max_exp (line 35) | def get_mx_max_exp(mx_dtype):
function get_p_contiguous_scale (line 47) | def get_p_contiguous_scale(hw_scale, data_p_size, p_offset=0):
function quantize_mx_golden (line 61) | def quantize_mx_golden(in_tensor, out_quantized_dtype, ocp_saturation = ...
function nc_matmul_mx_golden (line 112) | def nc_matmul_mx_golden(stationary_x4, moving_x4, stationary_scale, movi...
function dequantize_mx_golden (line 185) | def dequantize_mx_golden(mx_data_x4, quantized_dtype, mx_scale):
function generate_stabilized_mx_data (line 241) | def generate_stabilized_mx_data(quantized_dtype, shape, val_range=1.0):
function pack_mx_data_into_x4 (line 318) | def pack_mx_data_into_x4(mx_data):
function unpack_mx_data_from_x4 (line 345) | def unpack_mx_data_from_x4(packed_data, target_dtype):
FILE: nki/deep-dives/src/mxfp-matmul/mx_kernel_utils.py
function load_scales_scattered (line 18) | def load_scales_scattered(data_hbm, scale_hbm):
function load_tensor_helper (line 82) | def load_tensor_helper(stationary_hbm, moving_hbm):
function allocate_mx_tiles (line 104) | def allocate_mx_tiles(shape_unquantized, mx_dtype, alloc_scale: bool = T...
function copy_data_strided (line 136) | def copy_data_strided(stationary_hbm, moving_hbm, use_tensor_copy: bool ...
FILE: nki/deep-dives/src/mxfp-matmul/mx_kernels.py
function kernel_offline_quantized_mx_matmul (line 19) | def kernel_offline_quantized_mx_matmul(stationary_mx_data, stationary_mx...
function kernel_on_device_quantize_matmul_mx (line 83) | def kernel_on_device_quantize_matmul_mx(stationary_mx_data, stationary_m...
function kernel_copy_strided_quantize_matmul_mx (line 148) | def kernel_copy_strided_quantize_matmul_mx(stationary_hbm, moving_hbm, m...
function kernel_copy_strided_quantize_matmul_mx_packed_scale (line 210) | def kernel_copy_strided_quantize_matmul_mx_packed_scale(stationary_hbm, ...
FILE: nki/deep-dives/src/mxfp-matmul/mx_toplevel.py
function setup_compiler_workdir (line 24) | def setup_compiler_workdir(test_name):
function compare_and_print_results (line 37) | def compare_and_print_results(res, golden, rtol=5e-2, atol=5e-2):
function print_test_header (line 59) | def print_test_header(test_name):
function run_offline_quantized_matmul_mx_test (line 68) | def run_offline_quantized_matmul_mx_test(quantized_dtype):
function run_on_device_quantize_matmul_mx_test (line 105) | def run_on_device_quantize_matmul_mx_test(quantized_dtype_stationary, qu...
function run_copy_strided_test (line 154) | def run_copy_strided_test(quantized_dtype, use_tensor_copy: bool = True,...
FILE: nki/examples/average_pool2d/average_pool2d_jax.py
function jax_average_pool_2D (line 15) | def jax_average_pool_2D(in_tensor, pool_size):
FILE: nki/examples/average_pool2d/average_pool2d_nki_kernels.py
function tensor_avgpool_kernel (line 15) | def tensor_avgpool_kernel(in_tensor, pool_size):
function np_average_pool_2D (line 70) | def np_average_pool_2D(in_tensor, pool_size):
FILE: nki/examples/fused_mamba/mamba_nki_kernels.py
function mamba_v1 (line 18) | def mamba_v1(delta, u, A, B, C):
function mamba_v2 (line 108) | def mamba_v2(delta, u, A, B, C):
function mamba_v3 (line 198) | def mamba_v3(delta, u, A, B, C):
function parse_args (line 306) | def parse_args():
FILE: nki/examples/fused_mamba/mamba_torch.py
function associative_scan (line 18) | def associative_scan(deltaA, deltaB_u):
function mamba_layer (line 37) | def mamba_layer(delta, A, B, u, C):
function parse_args (line 56) | def parse_args():
FILE: nki/examples/getting_started_baremetal.py
function nki_tensor_add_kernel (line 9) | def nki_tensor_add_kernel(a_input, b_input):
FILE: nki/examples/getting_started_jax.py
function nki_tensor_add_kernel (line 5) | def nki_tensor_add_kernel(a_input, b_input):
FILE: nki/examples/getting_started_torch.py
function nki_tensor_add_kernel (line 5) | def nki_tensor_add_kernel(a_input, b_input):
FILE: nki/examples/index-case-1.py
function tensor_split_kernel_ (line 6) | def tensor_split_kernel_(in_tensor):
FILE: nki/examples/index-case-3.py
function tensor_maxpool_kernel_ (line 5) | def tensor_maxpool_kernel_(in_tensor, sz_pool):
FILE: nki/examples/layout-dynamic-loop.py
function tensor_exp_kernel_ (line 6) | def tensor_exp_kernel_(in_tensor):
FILE: nki/examples/layout-loop.py
function tensor_exp_kernel_ (line 5) | def tensor_exp_kernel_(in_tensor):
FILE: nki/examples/layout-pass.py
function tensor_exp_kernel_ (line 5) | def tensor_exp_kernel_(in_tensor):
FILE: nki/examples/layout-violation.py
function tensor_exp_kernel_ (line 6) | def tensor_exp_kernel_(in_tensor):
FILE: nki/examples/matrix_multiplication/matrix_multiplication_nki_kernels.py
function nki_matmul_basic_ (line 16) | def nki_matmul_basic_(lhsT, rhs):
function nki_matmul_tiled_ (line 74) | def nki_matmul_tiled_(lhsT, rhs):
function nki_matmul_hoist_load_ (line 146) | def nki_matmul_hoist_load_(lhsT, rhs):
function nki_matmul_block_free_dimension_ (line 231) | def nki_matmul_block_free_dimension_(lhsT, rhs):
function nki_matmul_fully_optimized_ (line 349) | def nki_matmul_fully_optimized_(
FILE: nki/examples/matrix_multiplication/matrix_multiplication_torch.py
function check_match (line 45) | def check_match(nki_func):
FILE: nki/examples/simulate/nki_simulate_example.py
function add_kernel (line 11) | def add_kernel(a_ptr, b_ptr):
FILE: nki/examples/tensor_addition/tensor_addition_nki_kernels.py
function nki_tensor_add (line 16) | def nki_tensor_add(a_input, b_input):
FILE: nki/examples/transpose2d/transpose2d_nki_kernels.py
function tensor_transpose2D_kernel_ (line 15) | def tensor_transpose2D_kernel_(in_tensor, shape2D):
FILE: nki/test/test_nki_isa_activation.py
function nki_activation (line 16) | def nki_activation(a_tensor, b_tensor, c_tensor):
class TestNkiIsaExamplesActivation (line 44) | class TestNkiIsaExamplesActivation(unittest.TestCase):
method test_activation (line 45) | def test_activation(self):
FILE: nki/test/test_nki_isa_affine_select.py
function nki_affine_select (line 16) | def nki_affine_select(a_tensor):
class TestNkiIsaExamplesAffineSelect (line 35) | class TestNkiIsaExamplesAffineSelect(unittest.TestCase):
method test_affine_select (line 36) | def test_affine_select(self):
FILE: nki/test/test_nki_isa_bn_stats.py
function nki_bn_stats_bn_aggr_1 (line 18) | def nki_bn_stats_bn_aggr_1(a_tensor):
function nki_bn_stats_bn_aggr_2 (line 42) | def nki_bn_stats_bn_aggr_2(b_tensor):
class TestNkiIsaExamplesBnStatsBnAggr (line 78) | class TestNkiIsaExamplesBnStatsBnAggr(unittest.TestCase):
method test_bn_stats_bn_aggr (line 79) | def test_bn_stats_bn_aggr(self):
FILE: nki/test/test_nki_isa_copypredicated.py
function nki_copy_predicated (line 18) | def nki_copy_predicated(predicate, on_true_tensor, on_false_tensor):
class TestNkiIsaExamplescopy_predicated (line 45) | class TestNkiIsaExamplescopy_predicated(unittest.TestCase):
method test_copy_predicated (line 46) | def test_copy_predicated(self):
FILE: nki/test/test_nki_isa_dma_copy.py
function nki_dma_copy (line 25) | def nki_dma_copy(a):
function nki_indirect_load_oob_err (line 40) | def nki_indirect_load_oob_err(in_tensor):
function nki_indirect_load_oob_error_negative (line 67) | def nki_indirect_load_oob_error_negative(in_tensor):
function nki_indirect_load_oob_skip (line 96) | def nki_indirect_load_oob_skip(in_tensor):
function nki_indirect_store_rmw (line 125) | def nki_indirect_store_rmw(in_tensor):
function nki_indirect_store_oob_err (line 153) | def nki_indirect_store_oob_err(in_tensor):
function nki_indirect_store_oob_err_negative (line 181) | def nki_indirect_store_oob_err_negative(in_tensor):
function nki_indirect_store_oob_skip (line 211) | def nki_indirect_store_oob_skip(in_tensor):
function nki_dma_copy_swdge (line 240) | def nki_dma_copy_swdge(in_tensor):
function nki_dma_copy_hwdge (line 259) | def nki_dma_copy_hwdge(in_tensor):
class TestNkiIsaExamplesTensorCopy (line 281) | class TestNkiIsaExamplesTensorCopy(unittest.TestCase):
method test_tensor_copy (line 282) | def test_tensor_copy(self):
method test_indirect_load_oob_err (line 291) | def test_indirect_load_oob_err(self):
method test_indirect_load_oob_err_negative (line 302) | def test_indirect_load_oob_err_negative(self):
method test_indirect_load_oob_skip (line 313) | def test_indirect_load_oob_skip(self):
method test_indirect_store_rmw (line 327) | def test_indirect_store_rmw(self):
method test_indirect_store_oob_err (line 340) | def test_indirect_store_oob_err(self):
method test_indirect_store_oob_err_negative (line 353) | def test_indirect_store_oob_err_negative(self):
method test_indirect_store_oob_skip (line 364) | def test_indirect_store_oob_skip(self):
method test_dma_copy_swdge (line 377) | def test_dma_copy_swdge(self):
method test_dma_copy_hwdge (line 383) | def test_dma_copy_hwdge(self):
FILE: nki/test/test_nki_isa_dma_transpose.py
function nki_dma_transpose_2d_hbm2sb (line 28) | def nki_dma_transpose_2d_hbm2sb(a):
function nki_dma_transpose_2d_sb2sb (line 40) | def nki_dma_transpose_2d_sb2sb(a):
function nki_dma_transpose_2d_hbm2sb_dge_xbar (line 53) | def nki_dma_transpose_2d_hbm2sb_dge_xbar(a):
function nki_dma_transpose_2d_sb2sb_dge_xbar (line 65) | def nki_dma_transpose_2d_sb2sb_dge_xbar(a):
function nki_dma_gather_transpose_3d_hbm2sb (line 78) | def nki_dma_gather_transpose_3d_hbm2sb(src_tensor, idx_tensor):
function nki_dma_gather_transpose_3d_sb2sb (line 99) | def nki_dma_gather_transpose_3d_sb2sb(src_tensor, idx_tensor):
class TestNkiIsaExamplesDmaTranspose (line 115) | class TestNkiIsaExamplesDmaTranspose(unittest.TestCase):
method test_dma_transpose_2d (line 116) | def test_dma_transpose_2d(self):
method test_dma_transpose_indirect (line 134) | def test_dma_transpose_indirect(self):
FILE: nki/test/test_nki_isa_dropout.py
function nki_dropout (line 16) | def nki_dropout(a_tensor, b_tensor):
function nki_dropout_scalar (line 36) | def nki_dropout_scalar(in_tensor):
class TestNkiIsaExamplesDropout (line 55) | class TestNkiIsaExamplesDropout(unittest.TestCase):
method test_dropout (line 56) | def test_dropout(self):
method test_dropout_scalar (line 67) | def test_dropout_scalar(self):
FILE: nki/test/test_nki_isa_iota.py
function nki_iota (line 17) | def nki_iota():
class TestNkiIsaExamplesIota (line 82) | class TestNkiIsaExamplesIota(unittest.TestCase):
method test_iota (line 83) | def test_iota(self):
FILE: nki/test/test_nki_isa_local_gather.py
function nki_local_gather (line 18) | def nki_local_gather(src_buffer, index, num_elem_per_idx, num_valid_indi...
class TestNkiIsaExamplesLocalGather (line 41) | class TestNkiIsaExamplesLocalGather(unittest.TestCase):
method test_local_gather (line 42) | def test_local_gather(self):
FILE: nki/test/test_nki_isa_max8.py
function nki_max8 (line 17) | def nki_max8():
class TestNkiIsaExamplesMax8 (line 34) | class TestNkiIsaExamplesMax8(unittest.TestCase):
method test_max8 (line 35) | def test_max8(self):
FILE: nki/test/test_nki_isa_memset.py
function nki_memset (line 22) | def nki_memset():
class TestNkiIsaExamplesMemset (line 38) | class TestNkiIsaExamplesMemset(unittest.TestCase):
method test_memset (line 39) | def test_memset(self):
FILE: nki/test/test_nki_isa_nc_find_index8.py
function nki_max_index8 (line 17) | def nki_max_index8():
class TestNkiIsaExamplesMaxIndex8 (line 43) | class TestNkiIsaExamplesMaxIndex8(unittest.TestCase):
method test_max_index8 (line 44) | def test_max_index8(self):
FILE: nki/test/test_nki_isa_nc_match_replace8.py
function nki_nc_match_replace8 (line 18) | def nki_nc_match_replace8():
function nki_nc_match_replace_indices8 (line 39) | def nki_nc_match_replace_indices8(in_tensor: nt.tensor, imm: np.float32):
function nki_nc_match_replace_indices8_mask (line 69) | def nki_nc_match_replace_indices8_mask(in_tensor: nt.tensor, imm: np.flo...
function nki_nc_match_replace_indices8_3d (line 102) | def nki_nc_match_replace_indices8_3d(data_tensor: nt.tensor):
function nki_nc_match_replace_indices8_3d_inplace (line 142) | def nki_nc_match_replace_indices8_3d_inplace(data_tensor: nt.tensor):
function match_and_get_index (line 179) | def match_and_get_index(data, vals):
function get_replaced_output_and_max_indices (line 194) | def get_replaced_output_and_max_indices(a, imm=0):
class TestNkiIsaExamplesMatchReplace8 (line 214) | class TestNkiIsaExamplesMatchReplace8(unittest.TestCase):
method test_nc_match_replace8 (line 215) | def test_nc_match_replace8(self):
method test_nc_match_replace_indices8 (line 230) | def test_nc_match_replace_indices8(self):
method test_nc_match_replace_indices8_mask (line 241) | def test_nc_match_replace_indices8_mask(self):
method test_nc_match_replace_indices8_3d (line 251) | def test_nc_match_replace_indices8_3d(self):
method test_nc_match_replace_indices8_3d_inplace (line 261) | def test_nc_match_replace_indices8_3d_inplace(self):
FILE: nki/test/test_nki_isa_nc_matmul.py
function nki_nc_matmul (line 22) | def nki_nc_matmul(a_tensor, b_tensor, d_tensor, e_tensor, g_tensor, h_te...
function nki_nc_matmul_double_row_gen3 (line 83) | def nki_nc_matmul_double_row_gen3(a_input, b_input):
class TestNkiIsaExamplesNcMatmul (line 104) | class TestNkiIsaExamplesNcMatmul(unittest.TestCase):
method test_nc_matmul (line 105) | def test_nc_matmul(self):
method test_double_row_gen3 (line 127) | def test_double_row_gen3(self):
FILE: nki/test/test_nki_isa_nc_stream_shuffle.py
function nki_nc_stream_shuffle (line 18) | def nki_nc_stream_shuffle(in_tensor):
function nki_nc_stream_shuffle_broadcast_partition (line 40) | def nki_nc_stream_shuffle_broadcast_partition(in_tensor):
function nki_nc_stream_shuffle_broadcast_mask (line 63) | def nki_nc_stream_shuffle_broadcast_mask(in_tensor):
class TestNkiIsaExamplesStreamShuffle (line 87) | class TestNkiIsaExamplesStreamShuffle(unittest.TestCase):
method test_stream_shuffle (line 88) | def test_stream_shuffle(self):
method test_broadcast_partition (line 94) | def test_broadcast_partition(self):
method test_broadcast_mask (line 100) | def test_broadcast_mask(self):
FILE: nki/test/test_nki_isa_nc_transpose.py
function nki_nc_transpose (line 23) | def nki_nc_transpose(a_tensor, b_tensor):
class TestNkiIsaExamplesSbTranspose (line 63) | class TestNkiIsaExamplesSbTranspose(unittest.TestCase):
method test_nc_transpose (line 64) | def test_nc_transpose(self):
FILE: nki/test/test_nki_isa_partition_reduce.py
function nki_par_reduce (line 23) | def nki_par_reduce(a_tensor, b_tensor):
function nki_par_reduce_nd_b (line 36) | def nki_par_reduce_nd_b(a_tensor, b_tensor):
class TestNkiIsaExamplesPartitionReduce (line 50) | class TestNkiIsaExamplesPartitionReduce(unittest.TestCase):
method test_par_reduce_nd (line 51) | def test_par_reduce_nd(self):
method test_par_reduce_nd_b (line 58) | def test_par_reduce_nd_b(self):
FILE: nki/test/test_nki_isa_range_select.py
function nki_range_select_example (line 19) | def nki_range_select_example(on_true, bound0, bound1, compare_op0, compa...
function nki_range_select_chaining (line 59) | def nki_range_select_chaining(on_true, bound0, bound1, compare_op0, comp...
class TestNkiIsaExamplesRangeSelect (line 150) | class TestNkiIsaExamplesRangeSelect(unittest.TestCase):
method test_range_select_example (line 151) | def test_range_select_example(self):
method test_range_select_chaining (line 177) | def test_range_select_chaining(self):
FILE: nki/test/test_nki_isa_reciprocal.py
function reciprocal_kernel (line 22) | def reciprocal_kernel(in_tensor):
class TestNkiExampleNisaReciprocal (line 35) | class TestNkiExampleNisaReciprocal(unittest.TestCase):
method test_nisa_reciprocal (line 36) | def test_nisa_reciprocal(self):
FILE: nki/test/test_nki_isa_reduce.py
function nki_reduce (line 22) | def nki_reduce(a_tensor):
class TestNkiIsaExamplesReduce (line 46) | class TestNkiIsaExamplesReduce(unittest.TestCase):
method test_reduce (line 47) | def test_reduce(self):
FILE: nki/test/test_nki_isa_select_reduce.py
function nki_select_reduce_basic (line 16) | def nki_select_reduce_basic(predicate_data, on_true_data):
function nki_select_reduce_with_reduction (line 49) | def nki_select_reduce_with_reduction(predicate_data, on_true_data, on_fa...
function nki_select_reduce_reverse_pred (line 90) | def nki_select_reduce_reverse_pred(predicate_data, on_true_data):
class TestNkiIsaExamplesSelectReduce (line 123) | class TestNkiIsaExamplesSelectReduce(unittest.TestCase):
method test_select_reduce_basic (line 124) | def test_select_reduce_basic(self):
method test_select_reduce_with_reduction (line 141) | def test_select_reduce_with_reduction(self):
method test_select_reduce_reverse_pred (line 158) | def test_select_reduce_reverse_pred(self):
FILE: nki/test/test_nki_isa_sequence_bounds.py
function nki_sequence_bounds (line 17) | def nki_sequence_bounds(segment_ids):
class TestNkiIsaExamplesSequenceBounds (line 44) | class TestNkiIsaExamplesSequenceBounds(unittest.TestCase):
method test_sequence_bounds (line 45) | def test_sequence_bounds(self):
FILE: nki/test/test_nki_isa_tensor_copy.py
function nki_tensor_copy (line 24) | def nki_tensor_copy(in_tensor):
class TestNkiIsaExamplesTensorCopy (line 40) | class TestNkiIsaExamplesTensorCopy(unittest.TestCase):
method test_tensor_copy (line 41) | def test_tensor_copy(self):
FILE: nki/test/test_nki_isa_tensor_scalar.py
function nki_tensor_scalar (line 22) | def nki_tensor_scalar(a_tensor, c_tensor, e_tensor, f_tensor):
class TestNkiIsaExamplesTensorScalar (line 79) | class TestNkiIsaExamplesTensorScalar(unittest.TestCase):
method test_tensor_scalar (line 80) | def test_tensor_scalar(self):
FILE: nki/test/test_nki_isa_tensor_scalar_cumulative.py
function nki_tensor_scalar_cumulative_scalar (line 15) | def nki_tensor_scalar_cumulative_scalar(
function nki_tensor_scalar_cumulative_vector (line 54) | def nki_tensor_scalar_cumulative_vector(
function nki_tensor_scalar_cumulative_chain (line 95) | def nki_tensor_scalar_cumulative_chain(
function nki_tensor_scan (line 145) | def nki_tensor_scan(src_data, op, initial):
class TestNkiIsaExamplesTensorScalarCumulative (line 176) | class TestNkiIsaExamplesTensorScalarCumulative(unittest.TestCase):
method test_tensor_scalar_cumulative_scalar1 (line 178) | def test_tensor_scalar_cumulative_scalar1(self):
method test_tensor_scalar_cumulative_scalar2 (line 192) | def test_tensor_scalar_cumulative_scalar2(self):
method test_tensor_scalar_cumulative_vector1 (line 206) | def test_tensor_scalar_cumulative_vector1(self):
method test_tensor_scalar_cumulative_vector2 (line 221) | def test_tensor_scalar_cumulative_vector2(self):
method test_tensor_scalar_cumulative_vector3 (line 236) | def test_tensor_scalar_cumulative_vector3(self):
method test_tensor_scalar_cumulative_load_reduce1 (line 251) | def test_tensor_scalar_cumulative_load_reduce1(self):
method test_tensor_scalar_cumulative_load_reduce2 (line 273) | def test_tensor_scalar_cumulative_load_reduce2(self):
method test_tensor_scalar_cumulative_load_reduce3 (line 295) | def test_tensor_scalar_cumulative_load_reduce3(self):
method test_tensor_scalar_cumulative_chain1 (line 317) | def test_tensor_scalar_cumulative_chain1(self):
method test_tensor_scalar_cumulative_chain2 (line 336) | def test_tensor_scalar_cumulative_chain2(self):
method test_tensor_scan (line 356) | def test_tensor_scan(self):
FILE: nki/test/test_nki_isa_tensor_tensor.py
function nki_tensor_tensor (line 23) | def nki_tensor_tensor(a_tensor, b_tensor):
class TestNkiIsaExamplesTensorTensor (line 43) | class TestNkiIsaExamplesTensorTensor(unittest.TestCase):
method test_tensor_tensor (line 44) | def test_tensor_tensor(self):
FILE: nki/test/test_nki_isa_tensor_tensor_scan.py
function nki_tensor_tensor_scan (line 23) | def nki_tensor_tensor_scan(a_tensor, b_tensor):
class TestNkiIsaExamplesTensorTensorScan (line 49) | class TestNkiIsaExamplesTensorTensorScan(unittest.TestCase):
method test_tensor_tensor_scan (line 50) | def test_tensor_tensor_scan(self):
FILE: nki/test/test_nki_mask.py
function nki_mask (line 22) | def nki_mask(in_tensor):
class TestNkiIsaExamplesMask (line 41) | class TestNkiIsaExamplesMask(unittest.TestCase):
method test_mask (line 42) | def test_mask(self):
FILE: nki/test/test_nki_memory_semantics.py
function simple_demo_kernel (line 8) | def simple_demo_kernel(a_ptr):
class TestNkiMemorySemantics (line 24) | class TestNkiMemorySemantics(unittest.TestCase):
method test_simulate_kernel (line 25) | def test_simulate_kernel(self):
FILE: nki/test/test_nki_nl_add.py
function add_tensors (line 20) | def add_tensors(a_tensor, b_tensor):
function add_tensor_scalar (line 34) | def add_tensor_scalar(a_tensor):
function add_broadcast_free_dim (line 48) | def add_broadcast_free_dim(a_tensor, b_tensor):
function add_broadcast_par_dim (line 62) | def add_broadcast_par_dim(a_tensor, b_tensor):
function add_broadcast_both_dims (line 76) | def add_broadcast_both_dims(a_tensor, b_tensor):
function add_broadcast_each_dims (line 90) | def add_broadcast_each_dims(a_tensor, b_tensor):
class TestNkiNlExampleAdd (line 103) | class TestNkiNlExampleAdd(unittest.TestCase):
method test_add (line 104) | def test_add(self):
method test_add_tensor_scalar (line 114) | def test_add_tensor_scalar(self):
method test_add_broadcast_free_dim (line 124) | def test_add_broadcast_free_dim(self):
method test_add_broadcast_par_dim (line 134) | def test_add_broadcast_par_dim(self):
method test_add_broadcast_both_dims (line 144) | def test_add_broadcast_both_dims(self):
method test_add_broadcast_each_dims (line 154) | def test_add_broadcast_each_dims(self):
FILE: nki/test/test_nki_nl_atomic_rmw.py
function atomic_rmw_indirect_indices (line 22) | def atomic_rmw_indirect_indices(in_tensor, indices_tensor, value_tensor):
class TestNkiExampleNlLoad (line 54) | class TestNkiExampleNlLoad(unittest.TestCase):
method test_atomic_rmw_indirect_indices (line 55) | def test_atomic_rmw_indirect_indices(self):
FILE: nki/test/test_nki_nl_broadcast.py
function test_nl_broadcast (line 20) | def test_nl_broadcast(in_tensor):
class TestNkiExampleNlBroadcast (line 43) | class TestNkiExampleNlBroadcast(unittest.TestCase):
method test_nl_broadcast_to (line 44) | def test_nl_broadcast_to(self):
FILE: nki/test/test_nki_nl_dslice.py
function example_kernel (line 17) | def example_kernel(in_tensor):
class TestNkiExampleNlLoad (line 32) | class TestNkiExampleNlLoad(unittest.TestCase):
method test_nl_load (line 33) | def test_nl_load(self):
FILE: nki/test/test_nki_nl_gather_flattened.py
function nki_gather_flattened (line 16) | def nki_gather_flattened():
class TestNkiExamplesGather (line 47) | class TestNkiExamplesGather(unittest.TestCase):
method test_gather_flattened (line 48) | def test_gather_flattened(self):
FILE: nki/test/test_nki_nl_load_store.py
function example_kernel (line 22) | def example_kernel(in_tensor, use_scalar=False):
function example_load_store_b (line 49) | def example_load_store_b(in_tensor):
class TestNkiExampleNlLoad (line 71) | class TestNkiExampleNlLoad(unittest.TestCase):
method test_nl_load (line 72) | def test_nl_load(self):
method test_nl_load_scalar (line 78) | def test_nl_load_scalar(self):
method test_load_store_3d (line 84) | def test_load_store_3d(self):
FILE: nki/test/test_nki_nl_load_store_indirect.py
function example_indirect_load_1 (line 24) | def example_indirect_load_1(data_tensor, idx_tensor):
function example_indirect_load_2 (line 48) | def example_indirect_load_2(data_tensor):
function example_indirect_save_1 (line 74) | def example_indirect_save_1(in_tensor, idx_tensor):
function example_indirect_save_2 (line 98) | def example_indirect_save_2(in_tensor):
class TestNkiExampleNlLoadStoreIndirect (line 122) | class TestNkiExampleNlLoadStoreIndirect(unittest.TestCase):
method test_indirect_load_1 (line 123) | def test_indirect_load_1(self):
method test_indirect_load_2 (line 131) | def test_indirect_load_2(self):
method test_indirect_save_1 (line 139) | def test_indirect_save_1(self):
method test_indirect_save_2 (line 146) | def test_indirect_save_2(self):
FILE: nki/test/test_nki_nl_load_transpose2d.py
function example_kernel_0 (line 23) | def example_kernel_0(in_tensor):
function example_kernel_1 (line 38) | def example_kernel_1(in_tensor):
class TestNkiExampleNlLoadTranspose2d (line 57) | class TestNkiExampleNlLoadTranspose2d(unittest.TestCase):
method test_dma_transpose_load_0 (line 58) | def test_dma_transpose_load_0(self):
method test_dma_transpose_load_1 (line 67) | def test_dma_transpose_load_1(self):
FILE: nki/test/test_nki_nl_mgrid.py
function example_kernel (line 22) | def example_kernel(in_tensor):
function example_kernel_1 (line 36) | def example_kernel_1(in_tensor):
class TestNkiExampleNlLoad (line 48) | class TestNkiExampleNlLoad(unittest.TestCase):
method test_nl_load (line 49) | def test_nl_load(self):
method test_nl_load_1 (line 56) | def test_nl_load_1(self):
FILE: nki/test/test_nki_simulate_kernel.py
function print_kernel (line 14) | def print_kernel(a_tensor):
class TestNkiIsaExamplesSimulateKernel (line 30) | class TestNkiIsaExamplesSimulateKernel(unittest.TestCase):
method test_simulate_kernel (line 31) | def test_simulate_kernel(self):
FILE: nki/test/test_nki_spmd_grid.py
function nki_spmd_kernel (line 15) | def nki_spmd_kernel(a):
class TestNkiIsaExamplesTensorCopy (line 31) | class TestNkiIsaExamplesTensorCopy(unittest.TestCase):
method test_spmd_grid (line 32) | def test_spmd_grid(self):
FILE: nki/test/test_psum_modulo_alloc.py
function num_elems (line 7) | def num_elems(shape):
function linearize (line 10) | def linearize(shape, indices):
function modulo_allocate_func (line 13) | def modulo_allocate_func(base, allocate_shape, scale):
function mod_alloc (line 24) | def mod_alloc(base_addr: int, *,
function allocated_loop_transpose (line 48) | def allocated_loop_transpose(a_ptr, tp_ptr):
class TestNkiPSUMModuloAllocation (line 84) | class TestNkiPSUMModuloAllocation(unittest.TestCase):
method test_simulate_kernel (line 85) | def test_simulate_kernel(self):
FILE: nki/test/test_sbuf_modulo_alloc.py
function num_elms (line 7) | def num_elms(shape):
function linearize (line 10) | def linearize(shape, indices):
function modulo_allocate_func (line 13) | def modulo_allocate_func(base, allocate_shape, scale):
function mod_alloc (line 24) | def mod_alloc(base_addr: int, *,
function allocated_loop_transpose (line 45) | def allocated_loop_transpose(a_ptr, tp_ptr):
class TestNkiSBUFModuloAllocation (line 82) | class TestNkiSBUFModuloAllocation(unittest.TestCase):
method test_simulate_kernel (line 83) | def test_simulate_kernel(self):
FILE: src/benchmark/tensorflow/distilbert-base-uncased-finetuned-sst-2-english_benchmark.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: src/benchmark/tensorflow/distilbert-base-uncased-finetuned-sst-2-english_compile.py
function get_batch (line 17) | def get_batch(tokenizer, sequence_length, batch_size):
FILE: src/examples/mxnet/data_parallel/benchmark_utils.py
class Results (line 6) | class Results():
method __init__ (line 8) | def __init__(self, batch_size, num_cores=1):
method add_result (line 15) | def add_result(self, latency_array, end_times, start_times):
method report (line 20) | def report(self, f, window_size=1):
FILE: src/examples/mxnet/data_parallel/parallel.py
function consumer (line 9) | def consumer(model_file, sample_input, input_queue, result_queue):
class NeuronSimpleDataParallel (line 35) | class NeuronSimpleDataParallel:
method __init__ (line 36) | def __init__(self, model_file, num_neuron_cores, sample_input):
method start_continuous_inference (line 60) | def start_continuous_inference(self):
method warmup (line 64) | def warmup(self, batch):
method infer (line 67) | def infer(self, batch):
method stop (line 72) | def stop(self):
method add_result (line 76) | def add_result(self, callback_fn):
method add_all_results (line 83) | def add_all_results(self, callback_fn):
FILE: src/examples/pytorch/bert_tutorial/bert_benchmark_utils.py
class BertTestDataset (line 11) | class BertTestDataset(torch.utils.data.Dataset):
method __init__ (line 14) | def __init__(self, tsv_file, tokenizer, max_length=128, transform=None):
method __len__ (line 33) | def __len__(self):
method __getitem__ (line 36) | def __getitem__(self, idx):
class BertResults (line 61) | class BertResults():
method __init__ (line 63) | def __init__(self, batch_size, num_cores=1):
method add_result (line 72) | def add_result(self, correct_count, inference_count, latency_array, en...
method report (line 80) | def report(self, f, window_size=1):
FILE: src/examples/pytorch/bert_tutorial/parallel.py
function consumer (line 9) | def consumer(model, input_queue):
class NeuronSimpleDataParallel (line 25) | class NeuronSimpleDataParallel():
method __init__ (line 27) | def __init__(self, model_file, num_neuron_cores, batch_size=1):
method eval (line 43) | def eval(self):
method train (line 47) | def train(self):
method start_continuous_inference (line 51) | def start_continuous_inference(self):
method infer (line 55) | def infer(self, batch, input_id, callback_fn):
method stop (line 58) | def stop(self):
FILE: src/examples/pytorch/byoc_sm_bert_tutorial/code/inference.py
function model_fn (line 11) | def model_fn(model_dir):
function input_fn (line 20) | def input_fn(serialized_input_data, content_type=JSON_CONTENT_TYPE):
function predict_fn (line 31) | def predict_fn(input_data, models):
function output_fn (line 51) | def output_fn(prediction_output, accept=JSON_CONTENT_TYPE):
FILE: src/examples/pytorch/libtorch_demo/bert_neuronx/detect_instance.py
function get_instance_type (line 18) | def get_instance_type() -> str:
function get_num_neuroncores (line 38) | def get_num_neuroncores(instance_type: Optional[str] = None) -> int:
function get_num_neuroncores_v3 (line 59) | def get_num_neuroncores_v3() -> int:
FILE: src/examples/pytorch/libtorch_demo/example_app/example_app.cpp
function Input (line 27) | Input get_input(const std::string& sentence_1, const std::string& senten...
function get_batch (line 73) | std::vector<torch::jit::IValue> get_batch(const std::vector<Input>& inputs)
function sanity_check (line 92) | int sanity_check(const std::string& model_filename)
function benchmark (line 142) | void benchmark(const std::string& model_filename, const std::vector<torc...
function main (line 165) | int main(int argc, char *argv[])
FILE: src/examples/pytorch/libtorch_demo/example_app/utils.cpp
function get_visible_cores_str (line 10) | std::string get_visible_cores_str(size_t num_neuron_cores, size_t cores_...
function get_uuid (line 17) | std::string get_uuid()
function get_model (line 48) | torch::jit::script::Module get_model(const std::string& filename)
FILE: src/examples/pytorch/libtorch_demo/tokenizers_binding/tokenizer_test.cpp
function main (line 12) | int main(int argc, char *argv[]) {
FILE: src/examples/pytorch/mnist_mlp/train_monitor.py
class MLP (line 16) | class MLP(nn.Module):
method __init__ (line 17) | def __init__(self, input_size = 28 * 28, output_size = 10, layers = [1...
method forward (line 23) | def forward(self, x):
function main (line 33) | def main():
FILE: src/examples/pytorch/mnist_mlp/train_tb.py
class MLP (line 18) | class MLP(nn.Module):
method __init__ (line 19) | def __init__(self, input_size = 28 * 28, output_size = 10, layers = [1...
method forward (line 25) | def forward(self, x):
function main (line 35) | def main():
FILE: src/examples/pytorch/neuronx_distributed/t5-inference/t5_model_layers.py
function prune_linear_layer (line 17) | def prune_linear_layer(layer: BaseParallelLinear, index: torch.LongTensor,
class ParallelAttention (line 55) | class ParallelAttention(T5Attention):
method __init__ (line 56) | def __init__(self, config: T5Config, has_relative_attention_bias=False):
method prune_heads (line 86) | def prune_heads(self, heads):
method compute_bias (line 102) | def compute_bias(self, query_length, key_length, device=None):
method forward (line 128) | def forward(
class ParallelSelfAttention (line 274) | class ParallelSelfAttention(T5LayerSelfAttention):
method __init__ (line 275) | def __init__(self, config, has_relative_attention_bias=False):
class ParallelCrossAttention (line 283) | class ParallelCrossAttention(T5LayerCrossAttention):
method __init__ (line 284) | def __init__(self, config):
class ParallelDenseActDense (line 291) | class ParallelDenseActDense(T5DenseActDense):
method __init__ (line 292) | def __init__(self, config: T5Config):
class ParallelDenseGatedActDense (line 300) | class ParallelDenseGatedActDense(T5DenseGatedActDense):
method __init__ (line 301) | def __init__(self, config: T5Config):
class ParallelFF (line 319) | class ParallelFF(T5LayerFF):
method __init__ (line 320) | def __init__(self, config: T5Config):
function load_pretrained_with_parallel_attn (line 331) | def load_pretrained_with_parallel_attn(model_name):
FILE: src/examples/pytorch/neuronx_distributed/t5-inference/t5_models.py
function get_wrapped_encoder (line 10) | def get_wrapped_encoder(max_length, num_beams, tp_degree, model_name):
function get_wrapped_decoder (line 28) | def get_wrapped_decoder(max_length, num_beams, tp_degree, model_name):
function parallel_trace_encoder (line 50) | def parallel_trace_encoder(model_name: str,
function parallel_trace_decoder (line 81) | def parallel_trace_decoder(model: T5ForConditionalGeneration,
FILE: src/examples/pytorch/neuronx_distributed/t5-inference/wrapper.py
class T5Wrapper (line 31) | class T5Wrapper(T5ForConditionalGeneration, NeuronGenerationMixin):
method _prepare_encoder_decoder_kwargs_for_generation (line 33) | def _prepare_encoder_decoder_kwargs_for_generation(
method prepare_inputs_for_generation (line 44) | def prepare_inputs_for_generation(
method _update_model_kwargs_for_xla_generation (line 75) | def _update_model_kwargs_for_xla_generation(
method _reorder_cache (line 107) | def _reorder_cache(self, past_key_values, beam_idx):
method infer (line 115) | def infer(self,
method parallel_infer (line 150) | def parallel_infer(self,
method forward (line 195) | def forward(
method beam_search (line 228) | def beam_search(
method greedy_search (line 400) | def greedy_search(
class EncoderWrapper (line 538) | class EncoderWrapper(torch.nn.Module):
method __init__ (line 543) | def __init__(self,
method forward (line 568) | def forward(self, input_ids, attention_mask):
class DecoderWrapper (line 631) | class DecoderWrapper(torch.nn.Module):
method __init__ (line 633) | def __init__(self,
method update_past (line 665) | def update_past(self, past_key_values):
method reorder_cache (line 676) | def reorder_cache(self, past_key_values, beam_idx):
method forward (line 681) | def forward(self,
FILE: src/examples/pytorch/torchserve/benchmark_bert.py
function one_thread (line 29) | def one_thread(pred, feed_data):
function current_performance (line 44) | def current_performance():
FILE: src/examples/pytorch/torchserve/handler_bert.py
class BertEmbeddingHandler (line 19) | class BertEmbeddingHandler(BaseHandler, ABC):
method __init__ (line 23) | def __init__(self):
method initialize (line 27) | def initialize(self, ctx):
method preprocess (line 50) | def preprocess(self, input_data):
method inference (line 82) | def inference(self, inputs):
method postprocess (line 106) | def postprocess(self, inference_output):
FILE: src/examples/pytorch/torchserve/handler_bert_neuronx.py
class BertEmbeddingHandler (line 19) | class BertEmbeddingHandler(BaseHandler, ABC):
method __init__ (line 23) | def __init__(self):
method initialize (line 27) | def initialize(self, ctx):
method preprocess (line 50) | def preprocess(self, input_data):
method inference (line 82) | def inference(self, inputs):
method postprocess (line 106) | def postprocess(self, inference_output):
FILE: src/examples/pytorch/torchserve/infer_bert.py
function worker_thread (line 18) | def worker_thread(worker_index):
FILE: src/examples/tensorflow/bert_demo/bert_client.py
function client (line 21) | def client():
function write_latencies (line 63) | def write_latencies():
FILE: src/examples/tensorflow/bert_demo/bert_model.py
function main (line 17) | def main():
class NeuronBERTMRPC (line 130) | class NeuronBERTMRPC:
method __init__ (line 132) | def __init__(self, bert_saved_model, dtype=tf.float16, batch_size=4, s...
method encoder (line 158) | def encoder(self, tensor, bias_tensor):
method fully_connected (line 169) | def fully_connected(self, input_tensor, layer_name):
method self_attention (line 184) | def self_attention(self, input_tensor, bias_tensor, layer_name):
method layer_norm (line 218) | def layer_norm(self, input_tensor, layer_name, force_float32=False):
method pooler_loss (line 233) | def pooler_loss(self, input_tensor):
method gelu_tanh (line 248) | def gelu_tanh(self, tensor):
method gelu_sigmoid (line 253) | def gelu_sigmoid(self, tensor):
FILE: src/examples/tensorflow/bert_demo/bert_model_server.py
function serve (line 17) | def serve():
FILE: src/examples/tensorflow/bert_demo/bert_no_model.py
function main (line 7) | def main():
FILE: src/examples/tensorflow/bert_demo/bert_server.py
class BERTService (line 36) | class BERTService(mrpc_pb2_grpc.mrpcServicer):
method __init__ (line 38) | def __init__(self, model_path, parallel, batch_size, bootstrap, vocab_...
method cleanup (line 104) | def cleanup(self):
method current_throughput (line 109) | def current_throughput(self):
method current_throughput_accuracy (line 124) | def current_throughput_accuracy(self):
method paraphrase (line 136) | def paraphrase(self, text_pair, context):
method put_input (line 147) | def put_input(self, text_a, text_b):
method process_input (line 155) | def process_input(self, idx):
method process_input_bootstrap (line 182) | def process_input_bootstrap(self, idx):
method get_output (line 195) | def get_output(self, iid):
function serve (line 202) | def serve():
FILE: src/examples/tensorflow/bert_demo/download_mrpc_data.py
function format_mrpc (line 10) | def format_mrpc(data_dir, path_to_data, path_to_dev_tsv):
function main (line 60) | def main(arguments):
FILE: src/examples/tensorflow/bert_demo/mrpc_feature.py
class InputExample (line 24) | class InputExample(object):
method __init__ (line 27) | def __init__(self, guid, text_a, text_b=None, label=None):
class PaddingInputExample (line 45) | class PaddingInputExample(object):
class InputFeatures (line 58) | class InputFeatures(object):
method __init__ (line 61) | def __init__(self,
function convert_single_example (line 74) | def convert_single_example(ex_index, example, label_list, max_seq_length,
function read_tsv (line 167) | def read_tsv(input_file, quotechar=None):
function create_examples (line 177) | def create_examples(lines, set_type):
function _truncate_seq_pair (line 195) | def _truncate_seq_pair(tokens_a, tokens_b, max_length):
function get_eval_model_feed_dict_list (line 212) | def get_eval_model_feed_dict_list(mrpc_tsv, vocab_txt):
function text_pair_to_model_feed_dict (line 239) | def text_pair_to_model_feed_dict(text_a, text_b, tokenizer):
FILE: src/examples/tensorflow/bert_demo/mrpc_pb2_grpc.py
class mrpcStub (line 14) | class mrpcStub(object):
method __init__ (line 18) | def __init__(self, channel):
class mrpcServicer (line 31) | class mrpcServicer(object):
method paraphrase (line 35) | def paraphrase(self, request, context):
function add_mrpcServicer_to_server (line 43) | def add_mrpcServicer_to_server(servicer, server):
FILE: src/examples/tensorflow/bert_demo/tokenization.py
function validate_case_matches_checkpoint (line 27) | def validate_case_matches_checkpoint(do_lower_case, init_checkpoint):
function convert_to_unicode (line 77) | def convert_to_unicode(text):
function printable_text (line 97) | def printable_text(text):
function load_vocab (line 120) | def load_vocab(vocab_file):
function convert_by_vocab (line 135) | def convert_by_vocab(vocab, items):
function convert_tokens_to_ids (line 143) | def convert_tokens_to_ids(vocab, tokens):
function convert_ids_to_tokens (line 147) | def convert_ids_to_tokens(inv_vocab, ids):
function whitespace_tokenize (line 151) | def whitespace_tokenize(text):
class FullTokenizer (line 160) | class FullTokenizer(object):
method __init__ (line 163) | def __init__(self, vocab_file, do_lower_case=True):
method tokenize (line 169) | def tokenize(self, text):
method convert_tokens_to_ids (line 177) | def convert_tokens_to_ids(self, tokens):
method convert_ids_to_tokens (line 180) | def convert_ids_to_tokens(self, ids):
class BasicTokenizer (line 184) | class BasicTokenizer(object):
method __init__ (line 187) | def __init__(self, do_lower_case=True):
method tokenize (line 195) | def tokenize(self, text):
method _run_strip_accents (line 219) | def _run_strip_accents(self, text):
method _run_split_on_punc (line 230) | def _run_split_on_punc(self, text):
method _tokenize_chinese_chars (line 250) | def _tokenize_chinese_chars(self, text):
method _is_chinese_char (line 263) | def _is_chinese_char(self, cp):
method _clean_text (line 285) | def _clean_text(self, text):
class WordpieceTokenizer (line 299) | class WordpieceTokenizer(object):
method __init__ (line 302) | def __init__(self, vocab, unk_token="[UNK]", max_input_chars_per_word=...
method tokenize (line 307) | def tokenize(self, text):
function _is_whitespace (line 361) | def _is_whitespace(char):
function _is_control (line 373) | def _is_control(char):
function _is_punctuation (line 385) | def _is_punctuation(char):
FILE: src/examples/tensorflow/keras_resnet50/fp32tofp16.py
function ConvertFP32ToOther (line 18) | def ConvertFP32ToOther(graphdef):
function load_graph (line 65) | def load_graph(model_file):
FILE: src/examples/tensorflow/keras_resnet50/infer_resnet50_keras.py
function pb_to_saved_model (line 23) | def pb_to_saved_model(pb_path, input_names, output_names, model_dir):
FILE: src/examples/tensorflow/keras_resnet50/infer_resnet50_keras_loadtest.py
function one_thread (line 69) | def one_thread(pred, model_feed_dict, index):
function current_throughput (line 84) | def current_throughput():
FILE: src/examples/tensorflow/keras_resnet50/optimize_for_inference.py
function clear_input (line 18) | def clear_input(node):
function replace_name (line 22) | def replace_name(node, name):
function replace_input (line 25) | def replace_input(node, input_name, new_name):
function swap_names (line 34) | def swap_names(node1, node2):
function get_const_node (line 39) | def get_const_node(const_node_name, const_by_name):
function get_const_ndarray (line 43) | def get_const_ndarray(const_node_name, const_by_name):
function adjust_bias_values (line 48) | def adjust_bias_values(bias_node, fbn_node, const_by_name):
function MoveBiasAddAfterFusedBatchNorm (line 58) | def MoveBiasAddAfterFusedBatchNorm(graphdef):
function FoldFusedBatchNorm (line 104) | def FoldFusedBatchNorm(graph_def):
function load_graph (line 127) | def load_graph(model_file):
FILE: src/examples/tensorflow/keras_resnet50/pb2sm_compile.py
function pb_to_saved_model (line 24) | def pb_to_saved_model(pb_path, input_names, output_names, model_dir):
FILE: src/examples/tensorflow/ssd300_demo/ssd300_detection.py
function main (line 13) | def main():
FILE: src/examples/tensorflow/ssd300_demo/ssd300_evaluation.py
function get_val_dataset (line 19) | def get_val_dataset(val_annotate, val_coco_root):
function main (line 26) | def main():
FILE: src/examples/tensorflow/ssd300_demo/ssd300_evaluation_client.py
function get_val_dataset (line 21) | def get_val_dataset(val_annotate, val_coco_root):
function main (line 28) | def main():
FILE: src/examples/tensorflow/ssd300_demo/ssd300_model.py
function decode_jpeg_resize (line 19) | def decode_jpeg_resize(input_tensor, image_size):
function preprocessor (line 35) | def preprocessor(input_tensor, image_size):
function tf_Conv2d (line 43) | def tf_Conv2d(input_tensor, module, first_conv=False):
function tf_BatchNorm2d (line 60) | def tf_BatchNorm2d(input_tensor, module):
function tf_MaxPool2d (line 69) | def tf_MaxPool2d(input_tensor, module):
function tf_Bottleneck (line 74) | def tf_Bottleneck(input_tensor, module):
function tf_SequentialBottleneck (line 88) | def tf_SequentialBottleneck(tensor, seq, resnet):
function tf_bbox_view (line 95) | def tf_bbox_view(detection_feed, modules, ndim):
function tf_feature_extractor (line 111) | def tf_feature_extractor(input_tensor, resnet):
function tf_box_predictor (line 128) | def tf_box_predictor(tensor, ssd300_torch):
function tf_ssd300 (line 148) | def tf_ssd300(input_tensor, ssd300_torch):
function scale_back_batch (line 155) | def scale_back_batch(bboxes_in, scores_in, scale_xy, scale_wh, dboxes_xy...
function select_nms_outputs (line 180) | def select_nms_outputs(input_tensors):
function postprocessor (line 184) | def postprocessor(ploc_ts, plabel_ts, bbox_scale_hw_ts, scale_xy, scale_...
class DefaultBoxes (line 215) | class DefaultBoxes(object):
method __init__ (line 217) | def __init__(self, fig_size, feat_size, steps, scales, aspect_ratios,
method scale_xy (line 262) | def scale_xy(self):
method scale_wh (line 266) | def scale_wh(self):
method __call__ (line 269) | def __call__(self, order="ltrb"):
function dboxes300_coco (line 274) | def dboxes300_coco():
function main (line 285) | def main():
FILE: src/examples/tensorflow/yolo_v3_demo/yolo_v3_coco_saved_model.py
class YOLOV3 (line 19) | class YOLOV3(object):
method __init__ (line 21) | def __init__(self, input_data, input_size, trainable):
method __build_nework (line 82) | def __build_nework(self, input_data):
method decode (line 129) | def decode(self, conv_output, anchors, stride, decoded_shape, input_si...
function darknet53 (line 169) | def darknet53(input_data, trainable):
function convolutional (line 204) | def convolutional(input_data, filters_shape, trainable, name, downsample...
function residual_block (line 239) | def residual_block(input_data, input_channel, filter_num1, filter_num2, ...
function upsample (line 250) | def upsample(input_data, name, method="deconv"):
function decode_jpeg_resize (line 267) | def decode_jpeg_resize(input_tensor, image_size):
function preprocessor (line 276) | def preprocessor(input_tensor, image_size):
function tf_postprocess_boxes (line 284) | def tf_postprocess_boxes(pred_xywh, pred_conf, pred_prob, org_img_shape,...
function convert_weights (line 313) | def convert_weights(org_weights_path, cur_weights_path, input_size):
function main (line 371) | def main():
FILE: src/examples/tensorflow/yolo_v4_demo/yolo_v4_coco_saved_model.py
function rename_weights (line 13) | def rename_weights(checkpoint):
function convert_pt_checkpoint_to_keras_h5 (line 671) | def convert_pt_checkpoint_to_keras_h5(state_dict):
class Mish (line 746) | class Mish(layers.Layer):
method __init__ (line 748) | def __init__(self):
method compute_output_shape (line 751) | def compute_output_shape(self, input_shape):
method call (line 754) | def call(self, x):
function conv2d_unit (line 758) | def conv2d_unit(x, filters, kernels, strides=1, padding='valid', bn=1, a...
function residual_block (line 775) | def residual_block(inputs, filters_1, filters_2):
function stack_residual_block (line 782) | def stack_residual_block(inputs, filters_1, filters_2, n):
function spp (line 789) | def spp(x):
function YOLOv4 (line 798) | def YOLOv4(inputs, num_classes, num_anchors, input_shape=(608, 608), ini...
function decode_jpeg_resize (line 1028) | def decode_jpeg_resize(input_tensor, image_size):
function preprocessor (line 1037) | def preprocessor(input_tensor, image_size):
function main (line 1045) | def main():
FILE: src/helperscripts/n2-helper.py
class manifest (line 8) | class manifest:
method __init__ (line 9) | def __init__(self, manifest_file):
method parse_manifest (line 14) | def parse_manifest(self):
method merge_release_packages (line 50) | def merge_release_packages(self):
method extract_major_minor_version (line 54) | def extract_major_minor_version(self, version):
method get_pip_packages_supporting_python_versions (line 58) | def get_pip_packages_supporting_python_versions(self, args):
method get_major_version (line 87) | def get_major_version(self, package_name, instance):
method generate_script (line 91) | def generate_script(self, args):
method install_dlami (line 194) | def install_dlami(self, args):
method install_neuron_dlami (line 213) | def install_neuron_dlami(self, args):
method jupyter_notebook (line 235) | def jupyter_notebook(self, args):
method install_and_enable_epel (line 258) | def install_and_enable_epel(self, args):
method config_neuron_repository (line 268) | def config_neuron_repository(self, args):
method get_repo (line 295) | def get_repo(self):
method update_os_packages (line 314) | def update_os_packages(self, args):
method install_os_headers (line 331) | def install_os_headers(self, args):
method install_git (line 349) | def install_git(self, args):
method install_neuron_driver (line 359) | def install_neuron_driver(self, args):
method install_neuron_runtime (line 423) | def install_neuron_runtime(self, args):
method install_efa_driver (line 485) | def install_efa_driver(self, args):
method install_neuron_system_tools (line 500) | def install_neuron_system_tools(self, args):
method install_extra_dependencies (line 550) | def install_extra_dependencies(self, args):
method set_python_venv (line 565) | def set_python_venv(self, args):
method activate_python_venv (line 631) | def activate_python_venv(self, args):
method set_pip_repository (line 655) | def set_pip_repository(self):
method install_aux (line 665) | def install_aux(self, args):
method install_neuron_compiler_and_framework (line 677) | def install_neuron_compiler_and_framework(self, args):
method install_neuron_compiler (line 819) | def install_neuron_compiler(self, args):
method install_neuron_framework (line 857) | def install_neuron_framework(self, args):
method get_latest_neuron_version_per_instance (line 964) | def get_latest_neuron_version_per_instance(self, instance):
method get_package_names (line 967) | def get_package_names(self, category, instance, neuron_version=None):
method get_package_version (line 979) | def get_package_version(self, category, neuron_version, name=None, fra...
method get_main_framework_version (line 996) | def get_main_framework_version(self, instance, framework, neuron_versi...
method list_packages (line 1013) | def list_packages(self, args):
method list_pyversions (line 1043) | def list_pyversions(self, args):
function cli_validate (line 1085) | def cli_validate(args):
function cli_parse_arguments (line 1096) | def cli_parse_arguments():
FILE: src/helperscripts/neuronsetuphelper.py
class neuron_release_info (line 11) | class neuron_release_info:
method __init__ (line 12) | def __init__(self):
function cli_parse_arguments (line 89) | def cli_parse_arguments():
function enumerate_release_manifest (line 116) | def enumerate_release_manifest(nr_setup, in_neuron_version):
function cli_validate (line 231) | def cli_validate(update,neuron_version,framework_version,is_latest_neuro...
function versiontuple (line 252) | def versiontuple(v):
function cli_list_cmd (line 262) | def cli_list_cmd(nr_setup, neuron_version, list):
function hlpr_print_config (line 303) | def hlpr_print_config(nr_setup, neuron_version):
function hlpr_build_pip_command (line 344) | def hlpr_build_pip_command(nr_setup, neuron_version, component,include_c...
function hlpr_pip_repos_setup (line 437) | def hlpr_pip_repos_setup():
function hlpr_pip_install_create_python_venv (line 447) | def hlpr_pip_install_create_python_venv(nr_setup, neuron_version):
function hlpr_pip_activate_python_venv (line 495) | def hlpr_pip_activate_python_venv(nr_setup, neuron_version):
function hlpr_framework_compiler_setup (line 511) | def hlpr_framework_compiler_setup(nr_setup, neuron_version, include_comp...
function hlpr_framework_dlami_activate (line 532) | def hlpr_framework_dlami_activate(nr_setup):
function hlpr_os_packages_update (line 555) | def hlpr_os_packages_update(nr_setup):
function hlpr_os_headers_update (line 571) | def hlpr_os_headers_update(nr_setup):
function hlpr_os_export_path (line 586) | def hlpr_os_export_path(nr_setup):
function hlpr_os_packages_first_setup (line 600) | def hlpr_os_packages_first_setup(nr_setup):
function hlpr_os_comp_setup_cmd (line 627) | def hlpr_os_comp_setup_cmd(nr_setup, neuron_version, comp,optional,pkg):
function hlpr_instructions (line 754) | def hlpr_instructions(nr_setup, neuron_version):
class neuron_setup_helper (line 925) | class neuron_setup_helper:
method __init__ (line 926) | def __init__(self, manifest_file,neuron_version):
method instructions (line 957) | def instructions(self,framework,action,framework_version,os,ami,mode):
FILE: src/libnrt/include/ndl/ndl.h
type ndl_nq_dev_t (line 18) | typedef enum NQ_DEV_TYPE {
type ndl_version_info_t (line 31) | typedef struct ndl_version_info {
type ndl_device_init_param_t (line 58) | typedef struct ndl_device_init_param {
type ndl_copy_buf_t (line 66) | typedef struct ndl_copy_buf {
type ndl_device_t (line 78) | typedef struct ndl_device {
type ndl_device_nc_t (line 95) | typedef struct ndl_device_nc {
type ndl_device_context_t (line 100) | typedef struct ndl_device_context {
type ndl_mem_info_t (line 104) | typedef struct ndl_mem_info {
type ndl_notification_context_t (line 116) | typedef struct ndl_notification_context {
type neuron_dma_eng_state (line 400) | struct neuron_dma_eng_state
type neuron_dma_queue_state (line 412) | struct neuron_dma_queue_state
type neuron_dma_queue_state (line 412) | struct neuron_dma_queue_state
type neuron_dma_queue_type (line 428) | enum neuron_dma_queue_type
type ndl_queue_init (line 446) | struct ndl_queue_init {
type ndl_queue_init_batch (line 458) | struct ndl_queue_init_batch {
type ndl_queue_init_batch (line 470) | struct ndl_queue_init_batch
type neuron_dma_queue_type (line 515) | enum neuron_dma_queue_type
type neuron_app_info (line 554) | struct neuron_app_info
type neuron_uuid (line 742) | struct neuron_uuid
type neuron_uuid (line 752) | struct neuron_uuid
type neuron_uuid (line 762) | struct neuron_uuid
type neuron_uuid (line 772) | struct neuron_uuid
type neuron_dm_block_type (line 900) | enum neuron_dm_block_type
type neuron_dm_resource_type (line 900) | enum neuron_dm_resource_type
type neuron_ioctl_mem_chunk_info (line 937) | struct neuron_ioctl_mem_chunk_info
type neuron_ioctl_nc_map (line 964) | struct neuron_ioctl_nc_map
type neuron_ioctl_nc_mapping_type (line 964) | enum neuron_ioctl_nc_mapping_type
type neuron_ultraserver_mode (line 1001) | enum neuron_ultraserver_mode
type neuron_ultraserver_mode (line 1012) | enum neuron_ultraserver_mode
type neuron_metrics_mode (line 1075) | enum neuron_metrics_mode
function nbitmap_test_bit (line 1092) | static inline uint32_t nbitmap_test_bit(uint32_t nr, uint64_t *addr)
function nbitmap_set_bit (line 1097) | static inline void nbitmap_set_bit(uint32_t nr, uint64_t *addr)
function nbitmap_ffs1 (line 1102) | static inline uint32_t nbitmap_ffs1(uint32_t nr, uint64_t *addr)
function nbitmap_popcount (line 1113) | static inline uint32_t nbitmap_popcount(uint32_t nr, uint64_t *addr)
function nbitmap_clr_bit (line 1122) | static inline void nbitmap_clr_bit(uint32_t nr, uint64_t *addr)
FILE: src/libnrt/include/ndl/neuron_driver_shared.h
type neuron_driver_feature_flag (line 11) | enum neuron_driver_feature_flag {
type neuron_pod_ctrl_req (line 38) | enum neuron_pod_ctrl_req {
type neuron_ultraserver_mode (line 45) | enum neuron_ultraserver_mode {
type neuron_metrics_mode (line 53) | enum neuron_metrics_mode {
type neuron_dma_queue_type (line 60) | enum neuron_dma_queue_type {
type neuron_cinit_state (line 66) | enum neuron_cinit_state {
type neuron_dma_eng_state (line 73) | struct neuron_dma_eng_state {
type neuron_dma_queue_state (line 81) | struct neuron_dma_queue_state {
type neuron_dma_h2t_ctx_handle_type (line 92) | enum neuron_dma_h2t_ctx_handle_type {
type NQ_DEVICE_TYPE (line 111) | enum NQ_DEVICE_TYPE {
type NQ_TYPE (line 117) | enum NQ_TYPE {
type neuron_dm_block_type (line 136) | enum neuron_dm_block_type {
type neuron_dm_resource_type (line 143) | enum neuron_dm_resource_type {
type neuron_uuid (line 150) | struct neuron_uuid {
type neuron_app_info (line 163) | struct neuron_app_info {
type nmetric_version_t (line 171) | typedef union nmetric_version {
type neuron_ioctl_mem_chunk_info (line 181) | struct neuron_ioctl_mem_chunk_info {
type neuron_ioctl_nc_mapping_type (line 190) | enum neuron_ioctl_nc_mapping_type {
type neuron_ioctl_nc_map_entry (line 193) | struct neuron_ioctl_nc_map_entry {
type neuron_ioctl_nc_map (line 197) | struct neuron_ioctl_nc_map {
type neuron_memcpy_batch_t (line 203) | typedef struct neuron_memcpy_batch {
type mem_alloc_category_t (line 216) | typedef enum {
type nds_header_t (line 361) | typedef struct nds_header {
FILE: src/libnrt/include/ndl/neuron_driver_shared_tensor_batch_op.h
type __u64 (line 10) | typedef __u64 nrt_tensor_batch_offset_t;
type __u64 (line 11) | typedef __u64 nrt_tensor_batch_size_t;
type nrt_tensor_batch_offset_t (line 14) | typedef uint64_t nrt_tensor_batch_offset_t;
type nrt_tensor_batch_size_t (line 15) | typedef uint64_t nrt_tensor_batch_size_t;
type nrt_tensor_batch_op_t (line 18) | typedef struct nrt_tensor_batch_op {
FILE: src/libnrt/include/nrt/ndebug_stream.h
type ndebug_stream_event_type_t (line 51) | typedef enum ndebug_stream_event_type {
type ndebug_stream_event_header_t (line 56) | typedef struct ndebug_stream_event_header {
type ndebug_stream_payload_debug_tensor_read_t (line 62) | typedef struct ndebug_stream_payload_debug_tensor_read {
FILE: src/libnrt/include/nrt/nds/neuron_ds.h
type nds_mem_usage_info_t (line 48) | typedef struct nds_mem_usage_info {
type nds_model_node_info_t (line 54) | typedef struct nds_model_node_info {
type nds_model_node_mem_usage_info_t (line 64) | typedef struct nds_model_node_mem_usage_info {
type nds_version_info_t (line 70) | typedef struct nds_version_info {
type nds_process_info_t (line 77) | typedef struct nds_process_info {
type nds_process_info_ext_t (line 86) | typedef struct nds_process_info_ext {
type nds_instance_t (line 90) | typedef struct nds_instance nds_instance_t;
type ndl_device_t (line 91) | typedef struct ndl_device ndl_device_t;
type feature_bitmap_bit_index_t (line 95) | typedef enum feature_bitmap_bit_index {
FILE: src/libnrt/include/nrt/nec.h
type ofi_comm_type_t (line 78) | typedef enum ofi_comm_type {
type enc_comm_type (line 86) | enum enc_comm_type {
type nec_pod_type_t (line 114) | typedef enum nec_pod_type {
type nec_pod_type_t (line 121) | typedef enum nec_pod_type {
type nec_pod_type_t (line 129) | typedef enum nec_pod_type {
type enc_comm (line 136) | struct enc_comm
type enc_channel (line 137) | struct enc_channel
type dma_addr_t (line 138) | typedef uint64_t dma_addr_t;
type enc_net_host_memory_index (line 140) | struct enc_net_host_memory_index {
type enc_net_host_memory (line 188) | struct enc_net_host_memory {
type enc_host_mem_t (line 199) | typedef struct enc_host_mem {
type enc_host_mem_shared_t (line 207) | typedef struct enc_host_mem_shared {
type enc_net_connector (line 215) | struct enc_net_connector {
type enc_pattern_t (line 238) | typedef enum enc_pattern {
type enc_net_connectivity_t (line 244) | typedef enum enc_net_connectivity {
type enc_channel (line 250) | struct enc_channel {
type enc_peer_info (line 277) | struct enc_peer_info {
type enc_topology_mode_t (line 284) | typedef enum enc_topology_mode {
type enc_comm_info (line 290) | struct enc_comm_info {
type enc_ring (line 313) | struct enc_ring {
type SEngine (line 328) | enum SEngine {
type enc_kangaring (line 337) | struct enc_kangaring {
type metaring_type_t (line 364) | typedef enum metaring_type {
type enc_alg_metaring (line 372) | struct enc_alg_metaring {
type enc_mesh_event_type_t (line 409) | typedef enum enc_mesh_event_type {
type enc_mesh_nbr_grp (line 482) | struct enc_mesh_nbr_grp {
type enc_mesh_event (line 487) | struct enc_mesh_event {
type enc_alg_mesh_type_t (line 494) | typedef enum enc_alg_mesh_type {
type enc_alg_mesh_subtype (line 506) | struct enc_alg_mesh_subtype {
type enc_alg_mesh (line 529) | struct enc_alg_mesh {
type enc_alg_hier (line 605) | struct enc_alg_hier {
type nccl_comm_info_t (line 642) | typedef struct nccl_comm_info {
type enc_nccl_comm_node_t (line 678) | typedef struct enc_nccl_comm_node {
type enc_proxy_histogram_config_t (line 706) | typedef struct enc_proxy_histogram_config {
type enc_neuron_device_info_t (line 715) | typedef struct enc_neuron_device_info {
type enc_comm (line 737) | struct enc_comm {
type enc_glb_comm (line 776) | struct enc_glb_comm {
type net_addr_mark_t (line 857) | typedef enum net_addr_mark {
type net_src_addr_t (line 863) | typedef struct net_src_addr {
type net_dest_addr_t (line 879) | typedef struct net_dest_addr {
type net_ops_info_t (line 891) | typedef struct net_ops_info {
type nec_version_info_t (line 942) | typedef struct nec_version_info {
FILE: src/libnrt/include/nrt/nrt.h
type nrt_model_t (line 25) | typedef struct nrt_model nrt_model_t;
type nrt_tensor_t (line 27) | typedef struct nrt_tensor nrt_tensor_t;
type nrt_cc_context_t (line 29) | typedef struct nrt_cc_context nrt_cc_context_t;
type nrt_tensor_placement_t (line 36) | typedef enum {
type nrt_framework_type_t (line 42) | typedef enum {
type nrt_op_type_t (line 86) | typedef enum nrt_op_type {
type nrt_dtype_t (line 94) | typedef enum nrt_dtype {
type nrt_cc_op_type_t (line 114) | typedef enum nrt_cc_op_type {
type nrt_instance_info_t (line 120) | typedef struct nrt_instance_info {
type nrt_tensor_set_t (line 240) | typedef void nrt_tensor_set_t;
type nrt_tensor_batch_t (line 353) | typedef struct nrt_tensor_batch {
type nrt_tensor_device_allocation_info_t (line 461) | typedef struct nrt_tensor_device_allocation_info {
type nrt_vnc_memory_stats_t (line 536) | typedef struct nrt_vnc_memory_stats {
type nrt_cc_comm_t (line 572) | typedef struct nrt_cc_comm {
type nrt_tensor_list_t (line 582) | typedef struct nrt_tensor_list {
FILE: src/libnrt/include/nrt/nrt_async.h
type nrta_xu_t (line 20) | typedef enum {
type nrta_seq_t (line 35) | typedef uint64_t nrta_seq_t;
type nrta_xu_id_t (line 36) | typedef uint16_t nrta_xu_id_t;
type nrta_error_t (line 44) | typedef struct nrta_error {
type nrta_error_tracker_t (line 51) | typedef struct nrta_error_tracker nrta_error_tracker_t;
FILE: src/libnrt/include/nrt/nrt_async_sendrecv.h
type nrt_async_sendrecv_comm_t (line 10) | typedef struct nrt_async_sendrecv_comm nrt_async_sendrecv_comm_t;
type nrt_async_sendrecv_request_t (line 11) | typedef struct nrt_async_sendrecv_request nrt_async_sendrecv_request_t;
FILE: src/libnrt/include/nrt/nrt_experimental.h
type nrt_tensor_usage_t (line 19) | typedef enum nrt_tensor_usage {
type nrt_tensor_info_t (line 26) | typedef struct nrt_tensor_info {
type nrt_tensor_info_array_t (line 35) | typedef struct nrt_tensor_info_array {
type nrt_ucode_img (line 92) | typedef struct nrt_ucode_img {
type nrt_ucode_info (line 97) | typedef struct nrt_ucode_info {
type nrt_model_info_t (line 162) | typedef struct nrt_model_info {
FILE: src/libnrt/include/nrt/nrt_profile.h
type nrt_profile_continuous_options_t (line 42) | typedef struct nrt_profile_continuous_options nrt_profile_continuous_opt...
type nrt_inspect_config_t (line 146) | typedef struct nrt_inspect_config nrt_inspect_config_t;
FILE: src/libnrt/include/nrt/nrt_status.h
type NRT_STATUS (line 14) | typedef enum {
FILE: src/libnrt/include/nrt/nrt_sys_trace.h
type nrt_sys_trace_config_t (line 21) | typedef struct nrt_sys_trace_config nrt_sys_trace_config_t;
type nrt_sys_trace_fetch_options_t (line 125) | typedef struct nrt_sys_trace_fetch_options nrt_sys_trace_fetch_options_t;
FILE: src/libnrt/include/nrt/nrt_version.h
type nrt_version_t (line 14) | typedef struct nrt_version {
FILE: src/neuron-gatherinfo/clear_params_tfpb.py
function zero_const (line 14) | def zero_const(node):
function ZeroAllConst (line 20) | def ZeroAllConst(graphdef):
function load_graph (line 41) | def load_graph(model_file):
FILE: src/neuron-gatherinfo/mx_neuron_check_model.py
class neuron_parser (line 9) | class neuron_parser:
method __init__ (line 10) | def __init__(self):
method get_neuron_supported (line 36) | def get_neuron_supported(self):
method get_tf_subgraph_types_names (line 43) | def get_tf_subgraph_types_names(self, node):
method load_tensorflow_model (line 53) | def load_tensorflow_model(self, path):
method get_mx_subgraph_types_names (line 68) | def get_mx_subgraph_types_names(self, node):
method load_mxnet_model (line 77) | def load_mxnet_model(self, path):
method get_cores_from_executable (line 97) | def get_cores_from_executable(executable):
method print_node_type_info (line 107) | def print_node_type_info(self):
method print_subgraph_ops (line 142) | def print_subgraph_ops(self, sg_nodetypes, sg_nodenames):
method print_neuron_node_info (line 163) | def print_neuron_node_info(self):
method print_neuron_support_stats (line 179) | def print_neuron_support_stats(self):
method print_common_desc (line 189) | def print_common_desc(self):
method run (line 198) | def run(self):
FILE: src/neuron-gatherinfo/neuron-gatherinfo.py
function get_os_version (line 135) | def get_os_version():
function get_files (line 159) | def get_files(*, basedir, matchfiles, verbose):
function dump_compiler_info (line 188) | def dump_compiler_info(*, outdir, location, allowmodel=False, addfldir=N...
function copy_stdout (line 232) | def copy_stdout(*, outdir, stdout, verbose):
function copy_syslog (line 251) | def copy_syslog(*, outdir, include_flag=False, verbose):
function dump_rt_info (line 299) | def dump_rt_info(*, location, verbose):
function allow_capture_of_files (line 315) | def allow_capture_of_files():
function add_additional_filters (line 333) | def add_additional_filters(filterfile):
function dump_miscinfo (line 351) | def dump_miscinfo(*, outdir, verbose):
function dump_proc_info (line 400) | def dump_proc_info(*, outdir, verbose):
function sanity_check (line 429) | def sanity_check(options):
function copy_files (line 475) | def copy_files(*, outdir, basedir, filelist, verbose):
function write_miscinfo (line 502) | def write_miscinfo(*, outdir, data):
function run_neuron_dump (line 521) | def run_neuron_dump(outdir, verbose):
function package_tarball (line 559) | def package_tarball(*, outdir, allowmodel, ccdir, verbose):
function add_cmdline_args (line 588) | def add_cmdline_args():
function main (line 655) | def main():
FILE: src/neuron-gatherinfo/tf_neuron_check_model.py
class neuron_parser (line 9) | class neuron_parser:
method __init__ (line 10) | def __init__(self):
method get_neuron_supported (line 35) | def get_neuron_supported(self):
method get_tf_subgraph_types_names (line 42) | def get_tf_subgraph_types_names(self, node):
method load_tensorflow_model (line 52) | def load_tensorflow_model(self, path):
method get_mx_subgraph_types_names (line 67) | def get_mx_subgraph_types_names(self, node):
method load_mxnet_model (line 76) | def load_mxnet_model(self, path):
method get_cores_from_executable (line 96) | def get_cores_from_executable(executable):
method print_node_type_info (line 106) | def print_node_type_info(self):
method print_subgraph_ops (line 141) | def print_subgraph_ops(self, sg_nodetypes, sg_nodenames):
method print_neuron_node_info (line 162) | def print_neuron_node_info(self):
method print_neuron_support_stats (line 178) | def print_neuron_support_stats(self):
method print_common_desc (line 188) | def print_common_desc(self):
method run (line 197) | def run(self):
FILE: src/neuronperf/conf.py
function run_apidoc (line 15) | def run_apidoc(app):
function setup (line 46) | def setup(app):
FILE: src/neuronperf/src/neuronperf/benchmarking.py
class Benchmarker (line 47) | class Benchmarker(threading.Thread):
method __init__ (line 60) | def __init__(
method _status (line 115) | def _status(self, status, error=None):
method next_input (line 124) | def next_input(self):
method prepare_inputs (line 133) | def prepare_inputs(self):
method load (line 153) | def load(self):
method warmup (line 158) | def warmup(self):
method setup (line 169) | def setup(self):
method infer (line 185) | def infer(self, worker_id) -> tuple:
method worker_thread (line 199) | def worker_thread(self, worker_id):
method run (line 226) | def run(self):
method stop (line 267) | def stop(self):
method results (line 279) | def results(self) -> dict:
class StatsThread (line 306) | class StatsThread(threading.Thread):
method __init__ (line 309) | def __init__(self, interval: float):
method run (line 316) | def run(self):
method join (line 323) | def join(self, **kwargs):
function _combine_results (line 328) | def _combine_results(results: List[dict]) -> dict:
function _get_num_workers (line 351) | def _get_num_workers(pipeline_size: int) -> int:
function get_instance_type (line 356) | def get_instance_type() -> str:
function _get_cost_per_hour (line 371) | def _get_cost_per_hour(instance_type: str) -> float:
function _get_max_neuroncores (line 386) | def _get_max_neuroncores(instance_type: str = None) -> int:
function _get_num_gpus (line 404) | def _get_num_gpus(instance_type: str = None) -> int:
function _get_num_devices (line 430) | def _get_num_devices(device_type: str, instance_type: str = None) -> int:
function _sanitize_inputs (line 443) | def _sanitize_inputs(inputs, batch_sizes: Union[int, List[int]], dataset...
function set_verbosity (line 474) | def set_verbosity(verbosity: int):
function compile (line 488) | def compile(
function run_benchmarker (line 673) | def run_benchmarker(benchmarker, duration, pipe=None):
function _run_benchmarker_new_interpreter (line 703) | def _run_benchmarker_new_interpreter(benchmarker, duration):
function _run_benchmarkers_multiprocess (line 797) | def _run_benchmarkers_multiprocess(
function _run_benchmarkers_multithreaded (line 838) | def _run_benchmarkers_multithreaded(
function run_benchmarkers (line 857) | def run_benchmarkers(
function _get_env_setup_fn (line 896) | def _get_env_setup_fn(benchmarker_id: int, benchmarker_config: dict, env...
function _get_setup_fn (line 958) | def _get_setup_fn(benchmarker_id: int, benchmarker_config: dict, setup_fn):
function _get_device_id (line 970) | def _get_device_id(benchmarker_id: int, benchmarker_config: dict):
function benchmark (line 987) | def benchmark(
FILE: src/neuronperf/src/neuronperf/cpu/cpu.py
class DummyModel (line 18) | class DummyModel:
method __call__ (line 19) | def __call__(self, x):
function benchmark (line 25) | def benchmark(model_class, inputs, *args, **kwargs):
FILE: src/neuronperf/src/neuronperf/logging.py
function _get_stream_handlers (line 15) | def _get_stream_handlers(level = logging.DEBUG):
FILE: src/neuronperf/src/neuronperf/model_index.py
function generate_id (line 31) | def generate_id(length: int = 8):
function generate_name (line 38) | def generate_name(model_name: str):
function _create (line 43) | def _create(model_name: str, compile_info: list) -> dict:
function create (line 56) | def create(
function delete (line 93) | def delete(filename: str):
function copy (line 117) | def copy(old_index: Union[str, dict], new_index: str, new_dir: str) -> str:
function move (line 138) | def move(old_index: str, new_index: str, new_dir: str) -> str:
function _sanitize (line 145) | def _sanitize(*model_indexes: Union[str, dict]) -> List[dict]:
function append (line 176) | def append(*model_indexes: Union[str, dict]) -> dict:
function save (line 216) | def save(model_index: dict, filename: str = None, root_dir=None) -> str:
function load (line 233) | def load(filename) -> dict:
function filter_configs (line 262) | def filter_configs(configs, filter_name, filter_values) -> List:
function filter (line 276) | def filter(index: Union[str, dict], **kwargs) -> dict:
FILE: src/neuronperf/src/neuronperf/mxnet/mxnet.py
class _MXNetModelWrapper (line 28) | class _MXNetModelWrapper:
method __init__ (line 29) | def __init__(self, device_id, sym, args, aux):
method __call__ (line 38) | def __call__(self, inputs):
function change_dir (line 66) | def change_dir(new_dir):
function _load_fn (line 75) | def _load_fn(model_filename, **kwargs):
function _compile_fn (line 81) | def _compile_fn(model, example_inputs, models_dir, model_name, **kwargs):
function compile (line 100) | def compile(model, inputs, *args, **kwargs):
function benchmark (line 104) | def benchmark(model_filename, inputs, *args, **kwargs):
FILE: src/neuronperf/src/neuronperf/reporting.py
function _validate_config (line 84) | def _validate_config(config):
function _validate_results (line 95) | def _validate_results(results):
function _get_report_name (line 106) | def _get_report_name(model_name: str) -> str:
function get_report (line 110) | def get_report(
function get_reports (line 243) | def get_reports(results, cost_per_hour: float = None) -> List[dict]:
function print_reports (line 267) | def print_reports(reports: List[dict], cols=PRINT_COLS, sort_by="through...
function write_csv (line 298) | def write_csv(reports: List[dict], filename: str = None, cols=CSV_COLS):
function write_json (line 325) | def write_json(reports: List[dict], filename: str = None):
FILE: src/neuronperf/src/neuronperf/scripts/run_benchmark_file.py
function main (line 6) | def main():
FILE: src/neuronperf/src/neuronperf/tensorflow/tensorflow.py
function _load_fn (line 22) | def _load_fn(model_file, **kwargs):
function _compile_fn (line 34) | def _compile_fn(model, inputs, models_dir, model_name, **kwargs):
function compile (line 64) | def compile(model, inputs, *args, **kwargs):
function benchmark (line 68) | def benchmark(model_filename, inputs, *args, **kwargs):
FILE: src/neuronperf/src/neuronperf/timing.py
function timestamp_convert (line 29) | def timestamp_convert(timestamps,
class Timer (line 47) | class Timer():
method __init__ (line 48) | def __init__(self,
method __enter__ (line 56) | def __enter__(self):
method __exit__ (line 59) | def __exit__(self, type, value, traceback):
method __delitem__ (line 62) | def __delitem__(self, index):
method __getitem__ (line 66) | def __getitem__(self, index):
method __iter__ (line 72) | def __iter__(self):
method __len__ (line 75) | def __len__(self):
method __str__ (line 78) | def __str__(self):
method start (line 81) | def start(self):
method stop (line 87) | def stop(self):
method next (line 92) | def next(self):
method reset (line 97) | def reset(self):
method insert (line 101) | def insert(self, timestamps: tuple, time_unit: str):
method start_timestamps (line 112) | def start_timestamps(self, time_unit: str = None):
method end_timestamps (line 116) | def end_timestamps(self, time_unit: str = None):
method timestamps (line 120) | def timestamps(self, time_unit: str = None):
method durations (line 128) | def durations(self, time_unit: str = None):
method total_duration (line 137) | def total_duration(self, time_unit: str = None):
method avg (line 147) | def avg(self, time_unit: str = None):
FILE: src/neuronperf/src/neuronperf/torch/torch.py
function _compile_fn (line 24) | def _compile_fn(model, example_inputs, models_dir, model_name, **kwargs):
function _load_fn (line 45) | def _load_fn(model_filename, **kwargs):
function _class_load_fn (line 53) | def _class_load_fn(model_class, **kwargs):
function compile (line 59) | def compile(model, inputs, *args, **kwargs):
function _get_dataset_loader_fn (line 64) | def _get_dataset_loader_fn(dataset, loop):
function benchmark (line 100) | def benchmark(model_filename, inputs, *args, dataset_inputs=False, loop_...
FILE: src/neuronperf/test/test_neuronperf.py
function test_timer (line 16) | def test_timer():
function test_timestamp_convert (line 52) | def test_timestamp_convert():
function test_model_index_create_from_file (line 65) | def test_model_index_create_from_file():
function test_model_index_create_delete_save_load (line 75) | def test_model_index_create_delete_save_load():
function test_model_index_copy (line 96) | def test_model_index_copy():
function test_model_index_copy_2 (line 125) | def test_model_index_copy_2():
function test_model_index_move (line 154) | def test_model_index_move():
function test_model_index_append (line 181) | def test_model_index_append():
function test_model_index_filter (line 198) | def test_model_index_filter():
function test_benchmarker (line 219) | def test_benchmarker():
function test_benchmark_multithread (line 234) | def test_benchmark_multithread():
function test_benchmark_multithread_2 (line 264) | def test_benchmark_multithread_2():
function test_benchmark_multiprocess (line 288) | def test_benchmark_multiprocess():
function test_benchmark_multiinterpreter (line 315) | def test_benchmark_multiinterpreter():
function test_reporting (line 337) | def test_reporting():
Copy disabled (too large)
Download .json
Condensed preview — 1636 files, each showing path, character count, and a content snippet. Download the .json file for the full structured content (11,317K chars).
[
{
"path": ".github/ISSUE_TEMPLATE/bug-report.yml",
"chars": 3159,
"preview": "---\nname: \"🐛 Bug Report\"\ndescription: Report a bug\ntitle: \"(short issue description)\"\nlabels: [bug, needs-triage]\nassign"
},
{
"path": ".github/ISSUE_TEMPLATE/config.yml",
"chars": 27,
"preview": "blank_issues_enabled: false"
},
{
"path": ".github/ISSUE_TEMPLATE/documentation.yml",
"chars": 572,
"preview": "---\nname: \"📕 Documentation Issue\"\ndescription: Report an issue in the documentation and Developer Guide\ntitle: \"(short i"
},
{
"path": ".github/ISSUE_TEMPLATE/feature-request.yml",
"chars": 1313,
"preview": "---\nname: 🚀 Feature Request\ndescription: Suggest an idea for this project\ntitle: \"(short issue description)\"\nlabels: [fe"
},
{
"path": ".github/pull_request_template.md",
"chars": 4545,
"preview": "**IMPORTANT!** _If this is a documentation PR for a specific release, this PR must go the corresponding release branch_ "
},
{
"path": ".github/stale_issue_mark_close_workflow.yml",
"chars": 705,
"preview": "name: Close inactive issues\non:\n schedule:\n - cron: \"30 1 * * *\"\n\njobs:\n close-issues:\n runs-on: ubuntu-latest\n "
},
{
"path": ".github/workflows/acknowledge-new-issue.yml",
"chars": 637,
"preview": "name: Acknowledge New Issue\n\non:\n issues:\n types: [opened]\n\npermissions:\n issues: write\n\njobs:\n acknowledge:\n r"
},
{
"path": ".github/workflows/auto-label-issues.yml",
"chars": 9783,
"preview": "# Auto-label issues based on content keywords\nname: auto-label-issues\n\non:\n issues:\n types: [opened]\n\njobs:\n auto-l"
},
{
"path": ".gitignore",
"chars": 166,
"preview": "_build/\n__pycache__/\n.venv/\n.DS_Store\nsrc/examples/pytorch/libtorch_demo.tar.gz\nsrc/neuronperf.tar.gz\n*-checkpoint.ipynb"
},
{
"path": ".readthedocs.yml",
"chars": 792,
"preview": "# .readthedocs.yml\n# Read the Docs configuration file\n# See https://docs.readthedocs.io/en/stable/config-file/v2.html fo"
},
{
"path": "CODEOWNERS",
"chars": 2242,
"preview": "# This file creates codeowners for the documentation. It will allow setting code reviewers for all Pull requests to merg"
},
{
"path": "CONTRIBUTING.md",
"chars": 7038,
"preview": "# Contributing Guidelines\n\nThank you for your interest in contributing to our project. Whether it's a bug report, new fe"
},
{
"path": "Dockerfile",
"chars": 402,
"preview": "FROM python:3.10-slim\n\nRUN apt-get update && apt-get install -y --no-install-recommends \\\n make enchant-2 git pandoc "
},
{
"path": "LICENSE-DOCUMENTATION",
"chars": 14845,
"preview": "*** Documentation:\n\nCreative Commons Attribution-ShareAlike 4.0 International Public License\n\nBy exercising the Licensed"
},
{
"path": "LICENSE-SAMPLECODE",
"chars": 931,
"preview": "Copyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.\n\nPermission is hereby granted, free of charge, t"
},
{
"path": "LICENSE-SUMMARY-DOCS-SAMPLES",
"chars": 363,
"preview": "*** Documentation and Sample Code:\n\nCopyright 2019 Amazon.com, Inc. or its affiliates. All Rights Reserved.\n\nThe documen"
},
{
"path": "Makefile",
"chars": 680,
"preview": "# Minimal makefile for Sphinx documentation\n#\n\n# You can set these variables from the command line, and also\n# from the "
},
{
"path": "README.md",
"chars": 1092,
"preview": "\n\n# AWS Neuron\n\n## Neuron SDK Overview\n\nAWS Neuron is a softwa"
},
{
"path": "_backup-setup/neuron-setup/multiframework/multi-framework-ubuntu22-neuron-dlami.rst",
"chars": 2376,
"preview": ".. _setup-ubuntu22-multi-framework-dlami:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup"
},
{
"path": "_backup-setup/neuron-setup/multiframework/multi-framework-ubuntu24-neuron-dlami.rst",
"chars": 2369,
"preview": ".. _setup-ubuntu24-multi-framework-dlami:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuron/amazon-linux/torch-neuron-al2-base-dlami.rst",
"chars": 2727,
"preview": ".. _setup-torch-neuron-al2-base-dlami:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-gu"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuron/amazon-linux/torch-neuron-al2-pytorch-dlami.rst",
"chars": 3253,
"preview": ".. _setup-torch-neuron-al2-pytorch-dlami:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuron/amazon-linux/torch-neuron-al2.rst",
"chars": 2290,
"preview": ".. _setup-torch-neuron-al2:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-index\n "
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuron/amazon-linux/torch-neuron-al2023.rst",
"chars": 2100,
"preview": ".. _setup-torch-neuron-al2023:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-inde"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuron/ubuntu/torch-neuron-ubuntu20-base-dlami.rst",
"chars": 2518,
"preview": ".. _setup-torch-neuron-u20-base-dlami:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-gu"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuron/ubuntu/torch-neuron-ubuntu20-pytorch-dlami.rst",
"chars": 3038,
"preview": ".. _setup-torch-neuron-u20-pytorch-dlami:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuron/ubuntu/torch-neuron-ubuntu20.rst",
"chars": 2068,
"preview": ".. _setup-torch-neuron-u20:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-index\n "
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuron/ubuntu/torch-neuron-ubuntu22.rst",
"chars": 2068,
"preview": ".. _setup-torch-neuron-u22:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-index\n "
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/amazon-linux/torch-neuronx-al2-base-dlami.rst",
"chars": 2802,
"preview": ".. _setup-torch-neuronx-al2-base-dlami:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-g"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/amazon-linux/torch-neuronx-al2-pytorch-dlami.rst",
"chars": 3511,
"preview": ".. _setup-torch-neuronx-al2-dlami-pytorch:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setu"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/amazon-linux/torch-neuronx-al2.rst",
"chars": 2361,
"preview": ".. _setup-torch-neuronx-al2:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-index\n"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/amazon-linux/torch-neuronx-al2023.rst",
"chars": 2178,
"preview": ".. _setup-torch-neuronx-al2023:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-ind"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu20-base-dlami.rst",
"chars": 2604,
"preview": ".. _setup-torch-neuronx-ubuntu20-base-dlami:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: se"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu20-pytorch-dlami.rst",
"chars": 3310,
"preview": ".. _setup-torch-neuronx-ubuntu20-dlami-pytorch:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link:"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu20.rst",
"chars": 2224,
"preview": ".. _setup-torch-neuronx-ubuntu20:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-i"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu22.rst",
"chars": 2213,
"preview": ".. _setup-torch-neuronx-ubuntu22:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-i"
},
{
"path": "_backup-setup/neuron-setup/pytorch/neuronx/ubuntu/torch-neuronx-ubuntu24.rst",
"chars": 2134,
"preview": ".. _setup-torch-neuronx-ubuntu24:\n\n.. card:: Select a Different Framework or Platform for Setup\n :link: setup-guide-i"
},
{
"path": "_content-types/conceptual-deep-dive.rst",
"chars": 3396,
"preview": ".. meta::\n :description: {short description here}\n :date_updated: {planned date of publication here}\n\n.. _{RST page "
},
{
"path": "_content-types/model-card.rst",
"chars": 3090,
"preview": ".. _unique-ref-id-here:\n\n.. meta::\n :description: AWS Neuron SDK model card for {Model Name}, version {version}. Over"
},
{
"path": "_content-types/procedural-how-to.rst",
"chars": 2627,
"preview": ".. meta::\n :description: {short description here}\n :date_updated: {planned date of publication here}\n\n.. _{RST page "
},
{
"path": "_content-types/procedural-tutorial.ipynb",
"chars": 6579,
"preview": "{\n \"cells\": [\n {\n \"cell_type\": \"raw\",\n \"metadata\": {\n \"raw_mimetype\": \"text/restructuredtext\",\n \"vscode\": {\n "
},
{
"path": "_content-types/reference-kernel-api.rst",
"chars": 2385,
"preview": ".. meta::\n :description: API reference for the {kernel-name} kernel included in the NKI Library .\n :date-modified:"
},
{
"path": "_content-types/release-notes-templates/compiler.rst",
"chars": 2560,
"preview": ".. _neuron-2-XX-0-compiler:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK compiler compo"
},
{
"path": "_content-types/release-notes-templates/containers.rst",
"chars": 2609,
"preview": ".. _neuron-2-XX-0-dlc:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK Deep Learning Conta"
},
{
"path": "_content-types/release-notes-templates/dlami.rst",
"chars": 2638,
"preview": ".. _neuron-2-XX-0-dlami:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK Deep Learning AWS"
},
{
"path": "_content-types/release-notes-templates/index.rst",
"chars": 7531,
"preview": ".. _neuron-2-XX-0-whatsnew:\n.. _latest-neuron-release:\n\n.. meta::\n :description: The official release notes for the AW"
},
{
"path": "_content-types/release-notes-templates/nki.rst",
"chars": 2591,
"preview": ".. _neuron-2-XX-0-nki:\n\n.. meta::\n :description: The official release notes for the AWS Neuron Kernel Interface (NKI) "
},
{
"path": "_content-types/release-notes-templates/nx-jax.rst",
"chars": 2606,
"preview": ".. _neuron-2-XX-0-jax:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK JAX support compone"
},
{
"path": "_content-types/release-notes-templates/nx-pytorch.rst",
"chars": 2295,
"preview": ".. _neuron-2-XX-0-pytorch:\n\n.. meta::\n :description: The official release notes for AWS Neuron SDK PyTorch support, ve"
},
{
"path": "_content-types/release-notes-templates/nxd-core.rst",
"chars": 2546,
"preview": ".. _neuron-2-XX-0-nxd-core:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK NxD Core compo"
},
{
"path": "_content-types/release-notes-templates/nxd-inference.rst",
"chars": 2581,
"preview": ".. _neuron-2-XX-0-nxd-inference:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK Transform"
},
{
"path": "_content-types/release-notes-templates/nxd-training.rst",
"chars": 2562,
"preview": ".. _neuron-2-XX-0-nxd-training:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK NxD Traini"
},
{
"path": "_content-types/release-notes-templates/runtime.rst",
"chars": 2556,
"preview": ".. _neuron-2-XX-0-runtime:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK Runtime compone"
},
{
"path": "_content-types/release-notes-templates/tools.rst",
"chars": 2564,
"preview": ".. _neuron-2-XX-0-tools:\n\n.. meta::\n :description: The official release notes for the AWS Neuron SDK Developer Tools c"
},
{
"path": "_ext/archive.py",
"chars": 922,
"preview": "# This file creates a downloadable archive from each directory listed in src_dirs.\n# You can modify or add additional ar"
},
{
"path": "_ext/df_tables.py",
"chars": 1540,
"preview": "import os\nfrom docutils.parsers.rst import Directive, directives\nfrom docutils.parsers.rst.directives.tables import CSVT"
},
{
"path": "_ext/local_documenter.py",
"chars": 1801,
"preview": "import os\nimport sys\n\nfrom sphinx.ext.autodoc import ModuleDocumenter, FunctionDocumenter\n\n\nclass LocalModuleDocumenter("
},
{
"path": "_ext/neuron_tag.py",
"chars": 28156,
"preview": "import os\n\nfrom docutils import nodes\nfrom docutils.statemachine import ViewList\n\nfrom sphinx.util.docutils import Sphin"
},
{
"path": "_ext/release-notes-automation-spec.md",
"chars": 11244,
"preview": "# Release Notes Review Automation Specification\n\n## Overview\n\nThis specification defines a GitHub Action that automatica"
},
{
"path": "_ext/release-notes-context.md",
"chars": 11948,
"preview": "# Release Notes Writing Guidelines\n\n## Core Principles\n\n### Answer Three Questions for Every Item\n\n- **What?** — What fe"
},
{
"path": "_ext/sphinx_plotly_directive.py",
"chars": 32464,
"preview": "\"\"\"\nCODE FROM: https://github.com/harupy/sphinx-plotly-directive\nLICENSE: MIT\n\nBased on: https://matplotlib.org/3.1.3/de"
},
{
"path": "_ext/symlink.py",
"chars": 1106,
"preview": "from docutils import nodes\nfrom docutils.parsers.rst import Directive, directives\n\nimport os, sys\n\ndef remove_symlink_ha"
},
{
"path": "_static/css/custom.css",
"chars": 782,
"preview": ".xxtable-smaller-font-size p, strong {\n font-size:0.9em; \n}\n\n.ablog-post-title p {\n font-size:0.9em; \n}\n\n.abl"
},
{
"path": "_static/css/custom.css.new",
"chars": 1229,
"preview": "\n.table-smaller-font-size p, strong {\n font-size: 90%; \n}\n\n \n\ntd, th , tr {\n white-space: normal !important;\n}\n"
},
{
"path": "_templates/recentposts.html",
"chars": 323,
"preview": "{% if ablog %}\n<h3>\n <a href=\"{{ pathto(ablog.blog_path) }}\">{{ gettext('Recent Posts') }}</a>\n</h3>\n<ul>\n {% set pcou"
},
{
"path": "_templates/search-field.html",
"chars": 4388,
"preview": "<form id=\"search-form\" class=\"bd-search d-flex align-items-center\" action=\"{{ pathto('search') }}\" method=\"get\"> \n <i c"
},
{
"path": "_templates/search-google.html",
"chars": 1421,
"preview": "{%- extends \"page.html\" %}\n{# Over-ride the body to be custom search structure we want #}\n{% block docs_body %}\n <div c"
},
{
"path": "_templates/search.html",
"chars": 1431,
"preview": "{%- extends \"page.html\" %}\n{# Over-ride the body to be custom search structure we want #}\n{% block docs_body %}\n <div c"
},
{
"path": "_utilities/JIRA_SETUP_QUICKSTART.md",
"chars": 2578,
"preview": "# Jira Integration Quick Start\n\n## Prerequisites Check\n\nRun these commands to verify you have everything installed:\n\n```"
},
{
"path": "_utilities/add_meta.py",
"chars": 6447,
"preview": "#!/usr/bin/env python3\n\"\"\"Add missing .. meta:: blocks with :description:, :keywords:, and :date-modified: to .rst files"
},
{
"path": "_utilities/audit_frameworks.py",
"chars": 15926,
"preview": "#!/usr/bin/env python3\n\"\"\"\nAudit script for the /frameworks directory of the AWS Neuron SDK documentation.\n\nDetects orph"
},
{
"path": "_utilities/check_urls.sh",
"chars": 1397,
"preview": "#!/bin/bash\n\n# Output file\noutput_file=\"url_check_results.txt\"\n\n# Initialize counters\ntotal=0\nworking=0\nnot_found=0\nothe"
},
{
"path": "_utilities/create_sitemap.py",
"chars": 1832,
"preview": "# v1.0 by dougeric 2025-09-30\n# Script to create sitemap.xml for Sphinx-generated docs; must be run at the root of the d"
},
{
"path": "_utilities/format_build_logs.py",
"chars": 18797,
"preview": "#!/usr/bin/env python3\n\"\"\"\nFormat Sphinx Build Logs\n\nThis script checks for Python 3.9 and pip, creates a virtual enviro"
},
{
"path": "_utilities/inject_archive_meta.py",
"chars": 5001,
"preview": "#!/usr/bin/env python3\n\"\"\"Inject noindex/nofollow meta directives and deprecation banners into archived .rst files.\"\"\"\n\n"
},
{
"path": "_utilities/metadata_schema.yaml",
"chars": 5731,
"preview": "# Metadata Schema for AWS Neuron SDK Setup Documentation\n# This schema defines the structured metadata fields used in se"
},
{
"path": "_utilities/migrate_setup_content.py",
"chars": 6403,
"preview": "#!/usr/bin/env python3\n\"\"\"\nSetup Content Migration Script\n\nMaps old setup file paths to new framework-first paths and ge"
},
{
"path": "_utilities/old-nki-apis.txt",
"chars": 10768,
"preview": "https://awsdocs-neuron.readthedocs-hosted.com/en/v2.26.1/nki/api/generated/nki.benchmark.html\nhttps://awsdocs-neuron.rea"
},
{
"path": "_utilities/setup_jira_token.sh",
"chars": 2363,
"preview": "#!/bin/bash\n# Setup script to fetch Jira API token from AWS Secrets Manager\n# and configure it for the Atlassian MCP ser"
},
{
"path": "about-neuron/amazonq-getstarted.rst",
"chars": 2510,
"preview": "\n.. image:: /images/q-logo.png\n :scale: 30%\n :alt: Amazon Q\n :align: left\n :target: https://aws."
},
{
"path": "about-neuron/announcements/index.rst",
"chars": 132,
"preview": ".. _announcements-main:\n\nAnnouncements\n=============\n\nThis page will be replaced by ABlog. It's here to make sure it's i"
},
{
"path": "about-neuron/announcements/neuron1.x/announce-eol-mx-before-1-5.rst",
"chars": 530,
"preview": ".. post:: May 01, 2023 01:00\n :language: en\n :tags: announce-eol mxnet-neuron\n\n.. _announce-eol-mxnet-before-1-5:\n"
},
{
"path": "about-neuron/announcements/neuron1.x/announce-eol-pt-1-5.rst",
"chars": 566,
"preview": ".. post:: Mar 25, 2022\n :language: en\n :tags: announce-eol torch-neuron\n\n.. _announce-eol-pt-1-5:\n\nAnnouncing end "
},
{
"path": "about-neuron/announcements/neuron1.x/announce-eol-pt-before-1-8.rst",
"chars": 552,
"preview": ".. post:: Nov 22, 2022\n :language: en\n :tags: announce-eol torch-neuron\n\n.. _announce-eol-pt-before-1-8:\n\nAnnounci"
},
{
"path": "about-neuron/announcements/neuron1.x/announce-eol-tf-before-2-5.rst",
"chars": 588,
"preview": ".. post:: Nov 22, 2022 01:00\n :language: en\n :tags: announce-eol tensorflow-neuron\n\n.. _announce-eol-tf-before-2-5"
},
{
"path": "about-neuron/announcements/neuron1.x/announce-eol-tf-before-2-7.rst",
"chars": 557,
"preview": ".. post:: May 01, 2023 01:00\n :language: en\n :tags: announce-eol tensorflow-neuron\n\n.. _announce-eol-tf-before-2-7"
},
{
"path": "about-neuron/announcements/neuron1.x/announcements.rst",
"chars": 11819,
"preview": ".. post:: Feb 17, 2022\n :language: en\n :tags: announcements\n\n.. _prev-announcements:\n\nPrevious Announcements\n====="
},
{
"path": "about-neuron/announcements/neuron1.x/eol-ncgs-env_2.rst",
"chars": 782,
"preview": ".. post:: Mar 25, 2022\n :language: en\n :tags: announce-eol\n\n\nAnnouncing end of support for ``NEURONCORE_GROUP_SIZE"
},
{
"path": "about-neuron/announcements/neuron1.x/eol-pt-15.rst",
"chars": 422,
"preview": ".. post:: Apr 29, 2022\n :language: en\n :tags: eol\n\n.. _eol-pt-15:\n\n\nEnd of support for torch-neuron version 1.5\n--"
},
{
"path": "about-neuron/announcements/neuron1.x/eol-tf-21-24.rst",
"chars": 524,
"preview": ".. post:: Mar 25, 2022\n :language: en\n :tags: eol\n\n.. _eol-tf-21-24:\n\nEnd of support for tensorflow-neuron version"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-component-change.rst",
"chars": 983,
"preview": ".. post:: December 21, 2023\n :language: en\n :tags: announce-name-change, neuron-component\n\n.. _announce-component-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-correction-neuron-driver-support-inf1.rst",
"chars": 1237,
"preview": ".. post:: March 12, 2026\n :language: en\n :tags: announce-correction-neuron-driver-inf1, neuron-driver-version, inf"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-deprecation-containers-rtd.rst",
"chars": 614,
"preview": ".. post:: December 20, 2023\n :language: en\n :tags: announce-deprecating-containers, runtime-rtd\n\n.. _announce-upda"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-deprecation-nxd-path-trace-api.rst",
"chars": 1140,
"preview": ".. post:: September 18, 2025\n :language: en\n :tags: announce-deprecation-nxd-path-trace-api, al2\n\n.. _announce-dep"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-deprecation-transformer-flag.rst",
"chars": 621,
"preview": ".. post:: September 15, 2023\n :language: en\n :tags: announce-end-of-support, transformer-flag \n\n.. _announce-end-o"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eol-megatron-lm.rst",
"chars": 761,
"preview": ".. post:: Aug 8, 2023\n :language: en\n :tags: announce-eol, trn1, trn1n\n\n.. _announce-eol-megatronlm:\n\nAnnouncing e"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eol-python-3-7.rst",
"chars": 507,
"preview": ".. post:: Jul 26, 2023 10:00\n :language: en\n :tags: announce-eol, python37\n\n.. _announce-eol-python37:\n\nAnnouncing"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eol-ubuntu-18.rst",
"chars": 473,
"preview": ".. post:: Jul 13, 2023 11:00\n :language: en\n :tags: announce-eol, ubuntu18\n\n.. _announce-eol-ubuntu18:\n\nAnnouncing"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-al2.rst",
"chars": 596,
"preview": ".. post:: June 28, 2024\n :language: en\n :tags: announce-eos-al2, al2\n\n.. _announce-eos-al2:\n\nAnnouncing end of sup"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-beta-pytorch-neuroncore-placement-apis.rst",
"chars": 885,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-no-longer-support-pytorch-neuroncore-placement\n\n.. _announ"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-bf16-vars.rst",
"chars": 840,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-no-longer-support-xla-env-vars\n\n.. _announce-eos-longer-su"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-block-dimension-nki.rst",
"chars": 735,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-eos-block-dimension-nki\n\n.. _announce-eos-block-dimension-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-dlami-ubuntu-22-04.rst",
"chars": 535,
"preview": ".. post:: December 18, 2025\n :language: en\n :tags: announce-eos-dlami-ubuntu-22-04\n\n.. _announce-eos-dlami-ubuntu-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-dlami.rst",
"chars": 2040,
"preview": ".. post:: April 24, 2024\n :language: en\n :tags: announce-eos-dlami, neuron-dlami\n\n.. _announce-eos-dlami:\n\nAnnounc"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-inf1-virtual-environments.rst",
"chars": 550,
"preview": ".. post:: December 18, 2025\n :language: en\n :tags: announce-eos-inf1-virtual-environments\n\n.. _announce-eos-inf1-v"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-jax-neuronx-nki-call.rst",
"chars": 741,
"preview": ".. post:: April 3, 2025\n :language: en\n :tags: announce-eos-jax-neuronx-features\n\n.. _announce-eos-jax-neuronx-fea"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-megatronlm-2-13.rst",
"chars": 654,
"preview": ".. post:: Aug 28, 2023\n :language: en\n :tags: announce-eos, trn1, trn1n\n\n.. _announce-eos-megatronlm:\n\nAWS Neuron "
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-mllama-checkpoint.rst",
"chars": 985,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: announce-eos-mllama-checkpoint\n\n.. _announce-eos-mllama-checkpoint:\n"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-multiframework-dlamis-inf1.rst",
"chars": 1180,
"preview": ".. post:: April 24, 2024\n :language: en\n :tags: announce-eos-dlamis-inf1, dlami-inf1\n\n.. _announce-update-multifra"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-nemo.rst",
"chars": 686,
"preview": ".. post:: April 3, 2025\n :language: en\n :tags: announce-eos-nemo-megatron\n\n.. _announce-eos-nnm:\n\nAnnouncing end o"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-neuron-det.rst",
"chars": 752,
"preview": ".. post:: December 20, 2024\n :language: en\n :tags: announce-eos-neuron-det\n\n.. _announce-eos-neuron-det:\n\nAnnounci"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-neuron-driver-support-inf1.rst",
"chars": 1106,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-eos-neuron-driver-2.21-version, neuron-driver-version, inf"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-neuron-profiler-2.rst",
"chars": 2354,
"preview": ".. post:: February 26, 2026\n :language: en\n :tags: announce-eos-neuron-profiler\n\n.. _announce-eos-neuron-profiler-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-neuron-profiler-v230.rst",
"chars": 2366,
"preview": ".. post:: March 31, 2026\n :language: en\n :tags: announce-eos-neuron-profiler\n\n.. _announce-eos-neuron-profiler-v23"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-neuron-profiler.rst",
"chars": 1998,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announce-eos-neuron-profiler\n\n.. _announce-eos-neuron-profiler:"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-neurondevice-version.rst",
"chars": 737,
"preview": ".. post:: June 28, 2024\n :language: en\n :tags: announce-eos-neuron-device-version, neuron-device-version\n\n.. _anno"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-neurondevice.rst",
"chars": 901,
"preview": ".. post:: June 28, 2024\n :language: en\n :tags: announce-eos-neuron-device, neuron-device\n\n.. _announce-eos-neurond"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-nxd-examples.rst",
"chars": 1750,
"preview": ".. post:: December 20, 2024\n :language: en\n :tags: announce-eos-nxd-examples\n\n.. _announce-eos-nxd-examples:\n\nAnno"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-nxdt-nxd-core-training.rst",
"chars": 1065,
"preview": ".. post:: February 26, 2026\n :language: en\n :tags: announce-eos-nxdt\n\n.. _announce-eos-nxdt-nxd-core-training:\n\nAn"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-probuf.rst",
"chars": 814,
"preview": ".. post:: June 28, 2024\n :language: en\n :tags: announce-eos-probuf, probuf\n\n.. _announce-eos-probuf319:\n\nAnnouncin"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-pt-versions.rst",
"chars": 566,
"preview": ".. post:: December 20, 2023\n :language: en\n :tags: announce-eos-pt, pt-versions\n\n.. _announce-eos_pytorch110:\n\nAnn"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-pt2.rst",
"chars": 584,
"preview": ".. post:: December 20, 2023\n :language: en\n :tags: announce-eos-pt-two, pt-versions-two\n\n.. _announce-eos_pytorch2"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-python38.rst",
"chars": 1222,
"preview": ".. post:: December 20, 2024\n :language: en\n :tags: announce-python-eos\n\n.. _announce-python-eos:\n\nAnnouncing end o"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-pytorch-1-1-3.rst",
"chars": 812,
"preview": ".. post:: December 20, 2024\n :language: en\n :tags: announce-eos-pytorch-version\n\n.. _announce-eos-pytorch-eos-113:"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-pytorch-1-9.rst",
"chars": 569,
"preview": ".. post:: August 28, 2023\n :language: en\n :tags: announce-eol, torch-neuron \n\n.. _announce-eol-pytorch19:\n\nAnnounc"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-pytorch-2-1.rst",
"chars": 651,
"preview": ".. post:: December 20, 2024\n :language: en\n :tags: announce-eos-pytorch-version\n\n.. _announce-eos-pytorch-2-1:\n\nAn"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-pytorch-2-7-2-8-v229.rst",
"chars": 653,
"preview": ".. post:: March 31, 2026\n :language: en\n :tags: announce-eos-pytorch-version\n\n.. _announce-eos-pytorch-2-7-2-8-v22"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-pytorch-2-7-2-8.rst",
"chars": 662,
"preview": ".. post:: February 26, 2026\n :language: en\n :tags: announce-eos-pytorch-version\n\n.. _announce-eos-pytorch-2-7-2-8:"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-pytorch-profiling-api.rst",
"chars": 1465,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announce-eos-pytorch-profling-api\n\n.. _announce-eos-pytorch-pro"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-tensorboard-tools.rst",
"chars": 855,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announce-eos-tensorboard-tools\n\n.. _announce-eos-tensorboard-to"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-tensorflow-2-8-9.rst",
"chars": 627,
"preview": ".. post:: April 3, 2025\n :language: en\n :tags: announce-tensorflow-versions-eos\n\n.. _announce-tfx-2-8-9-eos:\n\nAnno"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-tensorflow-inf2.rst",
"chars": 683,
"preview": ".. post:: February 26, 2026\n :language: en\n :tags: announce-eos-tensorflow\n\n.. _announce-eos-tensorflow-inf2:\n\nAnn"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-tensorflow1-x.rst",
"chars": 521,
"preview": ".. post:: June 28, 2024\n :language: en\n :tags: announce-tensorflow-eos, tf-versions-1-x\n\n.. _announce-tfx-eos:\n\nAn"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-torch-neuron.rst",
"chars": 587,
"preview": ".. post:: September 16, 2024\n :language: en\n :tags: announce-torch-neuron-eos, torch-neuron\n\n.. _announce-torch-ne"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-torch-neuronx-nki-jit.rst",
"chars": 640,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: announce-eos-torch-neuronx-nki-jit\n\n.. _announce-eos-torch-neuronx-n"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-u20-dlamis.rst",
"chars": 1113,
"preview": ".. post:: December 20, 2024\n :language: en\n :tags: announce-u20-dlami-dlc-eos\n\n.. _announce-u20-dlami-dlc-eos:\n\nAn"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-eos-xla-bf16.rst",
"chars": 858,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: announce-eos-xla-bf\n\n.. _announce-eos-xla-bf:\n\nAnnouncing end of sup"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-intent-eol-nemo-arg.rst",
"chars": 646,
"preview": ".. post:: Oct 26, 2023\n :language: en\n :tags: announce-intent-end-of-support-nemo-arg, nemo-arg\n\n.. _announce-inte"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-intent-eos-opt.rst",
"chars": 378,
"preview": ".. post:: Oct 26, 2023\n :language: en\n :tags: announce-intent-eos-opt, opt\n\n.. _announce-intent-eos-opt:\n\nAnnounci"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-intent-eos-pt-version.rst",
"chars": 657,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-eos-pt-two-five\n\n.. _announce-eos_pytorch25:\n\nAnnouncing E"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-intent-eos-pt2-6.rst",
"chars": 691,
"preview": ".. post:: September 18, 2025\n :language: en\n :tags: announce-eos-pt2-6\n\n.. _announce-eos_pt2-6:\n\nAnnouncing End of"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-intent-eos-tensorflow-tutorial-inf.rst",
"chars": 627,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-eos-tensorflow-tutorial\n\n.. _announce-eos-tensorflow-tutor"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-intent-eos-tnx.rst",
"chars": 684,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-eos-tnx\n\n.. _announce-eos-tnx:\n\nAnnouncing end of support "
},
{
"path": "about-neuron/announcements/neuron2.x/announce-intent-maintenance-tnx.rst",
"chars": 640,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: announce-transformers-neuronx-maintenance, tnx\n\n.. _announce-tnx-mai"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-maintenance-mxnet.rst",
"chars": 578,
"preview": ".. post:: June 28, 2024\n :language: en\n :tags: announce-mxnet-maintenance, mxnet\n\n.. _announce-mxnet-maintenance:\n"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-maintenance-nxdi-nxd-core-inference.rst",
"chars": 768,
"preview": ".. post:: March 31, 2026\n :language: en\n :tags: announce-maintenance-nxdi\n\n.. _announce-maintenance-nxdi-nxd-core-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-maintenance-nxdt-nxd-core-training.rst",
"chars": 992,
"preview": ".. post:: March 31, 2026\n :language: en\n :tags: announce-maintenance-nxdt\n\n.. _announce-maintenance-nxdt-nxd-core-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-maintenance-tf.rst",
"chars": 562,
"preview": ".. post:: April 1, 2024\n :language: en\n :tags: announce-tensorflow-maintenance, tf-versions\n\n.. _announce-tfx-main"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-moving-samples.rst",
"chars": 935,
"preview": ".. post:: December 20, 2023\n :language: en\n :tags: announce-moving-nxd-samples, nxd-samples\n\n.. _announce-moving-s"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-nki-library-namespace-changes-2-28.rst",
"chars": 688,
"preview": ".. post:: February 26, 2026\n :language: en\n :tags: announce-nki-library-changes\n\n.. _announce-nki-library-namespac"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-nki-namespace-migration.rst",
"chars": 714,
"preview": ".. post:: March 31, 2026\n :language: en\n :tags: announce-nki-namespace\n\n.. _announce-nki-namespace-migration:\n\nAnn"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-longer-support-neuron-det.rst",
"chars": 780,
"preview": ".. post:: April 3, 2025\n :language: en\n :tags: announce-no-longer-support-neuron-det\n\n.. _announce-no-longer-suppo"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-longer-support-nxd-examples.rst",
"chars": 1768,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: announce-eol-nxd-examples\n\n.. _announce-eol-nxd-examples:\n\nAnnouncin"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-longer-support-pytorch-113.rst",
"chars": 831,
"preview": ".. post:: April 3, 2025\n :language: en\n :tags: announce-no-longer-support-pytorch-version\n\n.. _announce-no-longer-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-longer-support-pytorch-2-1.rst",
"chars": 689,
"preview": ".. post:: April 3, 2025\n :language: en\n :tags: announce-no-longer-support-pytorch-version\n\n.. _announce-no-longer-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-longer-support-pytorch-2-7-2-8.rst",
"chars": 676,
"preview": ".. post:: March 30, 2026\n :language: en\n :tags: announce-no-longer-support-pytorch-version\n\n.. _announce-no-longer"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-longer-support-tensorflow-inf2.rst",
"chars": 568,
"preview": ".. post:: March 30, 2026\n :language: en\n :tags: announce-no-longer-support-tensorflow\n\n.. _announce-no-longer-supp"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-longer-support-u20-dlc-dlami.rst",
"chars": 1040,
"preview": ".. post:: April 3, 2025\n :language: en\n :tags: announce-u20-dlami-dlc-no-longer-support\n\n.. _announce-u20-dlami-dl"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-al2.rst",
"chars": 2879,
"preview": ".. post:: September 16, 2024\n :language: en\n :tags: end-support-al2\n\n.. _eos-al2:\n\nNeuron Runtime no longer suppor"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-device-version.rst",
"chars": 862,
"preview": ".. post:: September 16, 2024\n :language: en\n :tags: eos-neuron-device, neuron-device-\n\n.. _eos-neurondevice:\n\n'neu"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-jax-neuronx-nki-call.rst",
"chars": 654,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: \n\n.. _announce-eos-jax-neuronx-features:\n\nNeuron no longer supports "
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-llama3-2-checkpoint.rst",
"chars": 988,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-no-longer-support-llama-checkpoint\n\n.. _announce-no-longer"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-nemo-megatron.rst",
"chars": 679,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: announce-no-support-nemo-megatron\n\n.. _announce-no-support-nemo-mega"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-neurondevice.rst",
"chars": 688,
"preview": ".. post:: September 16, 2024\n :language: en\n :tags: eos-neuron-device-version, neuron-device-version\n\n.. _eos-neur"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-nki-jit-torch.rst",
"chars": 646,
"preview": ".. post:: June 24, 2025\n :language: en\n :tags: announce-no-longer-support-nki-jit\n\n.. _announce-no-longer-support-"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-tensorboard-plugin.rst",
"chars": 728,
"preview": ".. post:: February 26, 2026\n :language: en\n :tags: announce-no-support-tensorboard\n\n.. _announce-no-support-tensor"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-tensorflow1-x.rst",
"chars": 525,
"preview": ".. post:: September 16, 2024\n :language: en\n :tags: no-support-tensorflow-eos, tf-versions-1-x-no-support\n\n.. _ann"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-tensorflow2-10.rst",
"chars": 919,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announce-no-support-tensorflow2-10\n\n.. _announce-no-support-ten"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-tf-versions.rst",
"chars": 588,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: announce-no-support-tensorflow-eos\n\n.. _announce-no-support-tensorfl"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-torch-neuron-versions.rst",
"chars": 548,
"preview": ".. post:: December 20, 2024\n :language: en\n :tags: announce-no-support-torch-neuron\n\n.. _announce-no-support-torch"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-ubuntu-20-base.rst",
"chars": 732,
"preview": ".. post:: May 15, 2025\n :language: en\n :tags: announce-u20-base-no-support\n\n.. _announce-u20-base-no-support:\n\nNeu"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-no-support-vllm-v0.rst",
"chars": 897,
"preview": ".. post:: February 26, 2026\n :language: en\n :tags: announce-no-support-vllm\n\n.. _announce-no-support-vllm-v0:\n\nNeu"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-nxdi-changes.rst",
"chars": 1020,
"preview": ".. post:: December 19, 2025\n :language: en\n :tags: announce-nxdi-changes\n\n.. _announce-nxdi-changes:\n\nAnnouncing c"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-package-change.rst",
"chars": 840,
"preview": ".. post:: September 16, 2024\n :language: en\n :tags: announce-nxdcore, neuron-component-nxdcore\n\n.. _announce-compo"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-python38-no-longer-support.rst",
"chars": 1300,
"preview": ".. post:: April 3, 2025\n :language: en\n :tags: announce-python-version-no-longer-support\n\n.. _announce-python-no-l"
},
{
"path": "about-neuron/announcements/neuron2.x/announce-transition-pytorch-trainium.rst",
"chars": 1313,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announce-transition-pytorch-trainium\n\n.. _announce-transition-p"
},
{
"path": "about-neuron/announcements/neuron2.x/announcement-end-of-support-neuronxcc-nki.rst",
"chars": 810,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announcement-end-of-support-neuronxcc-nki\n\n.. _announcement-end"
},
{
"path": "about-neuron/announcements/neuron2.x/announcement-end-of-support-nxdt-nxd-core.rst",
"chars": 1104,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announcement-end-of-support-nxdt-nxd-core\n\n.. _announcement-end"
},
{
"path": "about-neuron/announcements/neuron2.x/announcement-end-of-support-parallel-model-trace.rst",
"chars": 867,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announcement-end-of-support-parallel-model-trace\n\n.. _announcem"
},
{
"path": "about-neuron/announcements/neuron2.x/announcement-end-of-support-pytorch-2-6.rst",
"chars": 690,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announcement-end-of-support-pytorch-2-6\n\n.. _announcement-end-o"
},
{
"path": "about-neuron/announcements/neuron2.x/announcement-end-of-support-vllm-v0.rst",
"chars": 1269,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announcement-end-of-support-vllm-v0\n\n.. _announcement-end-of-su"
},
{
"path": "about-neuron/announcements/neuron2.x/announcement-nki-library-kernel-migration.rst",
"chars": 843,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announcement-nki-library-kernel-migration\n\n.. _announcement-nki"
},
{
"path": "about-neuron/announcements/neuron2.x/announcement-nki-library-namespace-changes.rst",
"chars": 927,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announcement-nki-library-namespace-changes\n\n.. _announcement-nk"
},
{
"path": "about-neuron/announcements/neuron2.x/announcement-python-3-9-eol.rst",
"chars": 728,
"preview": ".. post:: December 16, 2025\n :language: en\n :tags: announcement-python-3-9-eol\n\n.. _announcement-python-3-9-eol:\n\n"
},
{
"path": "about-neuron/announcements/neuron2.x/dlami-neuron-2.10.rst",
"chars": 1106,
"preview": ".. post:: May 02, 2023 11:00\n :language: en\n :tags: dlami, pytorch, trn1, inf2, inf1\n\n.. _announce-dlc-sm-neuron-2"
},
{
"path": "about-neuron/announcements/neuron2.x/dlami-neuron-2.12.rst",
"chars": 1356,
"preview": ".. post:: July 26, 2023 11:00\n :language: en\n :tags: dlami, pytorch, trn1, inf2, inf1\n\n.. _announce-dlami-neuron-2"
},
{
"path": "about-neuron/announcements/neuron2.x/dlami-pytorch-introduce.rst",
"chars": 766,
"preview": ".. post:: Nov 02, 2022 00:01\n :language: en\n :tags: dlami, pytorch\n\n.. _announce-dlami-neuron-pytorch:\n\nIntroducin"
},
{
"path": "about-neuron/announcements/neuron2.x/end-of-support-pt2.rst",
"chars": 415,
"preview": ".. post:: February 2, 2024\n :language: en\n :tags: eos-pt-two, pt-two\n\n.. _eos_pytorch2:\n\nPyTorch NeuronX version 2"
},
{
"path": "about-neuron/announcements/neuron2.x/github-changes.rst",
"chars": 1283,
"preview": ".. post:: Oct 10, 2022 02:00\n :language: en\n :tags: github\n\n.. _announce-aws-neuron-github-org:\n\nIntroducing New N"
},
{
"path": "about-neuron/announcements/neuron2.x/gpg-expiration.rst",
"chars": 783,
"preview": ".. post:: Nov 10, 2022 00:01\n :language: en\n :tags: dlami, pytorch\n\n.. _announce-dlami-neuron-pytorch:\n\nNeuron GPG"
},
{
"path": "about-neuron/announcements/neuron2.x/neuron-rtd-eol.rst",
"chars": 1225,
"preview": ".. post:: Oct 10, 2022 01:00\n :language: en\n :tags: eol, neuron2.x\n\n.. _announce-neuron-rtd-eol:\n\nAnnouncing Neuro"
},
{
"path": "about-neuron/announcements/neuron2.x/neuron2-intro.rst",
"chars": 1246,
"preview": ".. post:: Oct 10, 2022 04:00\n :language: en\n :tags: neuron2.x\n\n.. _neuron2-intro:\n\nIntroducing the first release o"
},
{
"path": "about-neuron/announcements/neuron2.x/neuron230-packages-changes.rst",
"chars": 4929,
"preview": ".. post:: Oct 10, 2022 03:00\n :language: en\n :tags: neuron2.x\n\n.. _neuron-packages-changes:\n\nIntroducing Packaging"
},
{
"path": "about-neuron/announcements/neuron2.x/neuron250-packages-changes.rst",
"chars": 1544,
"preview": ".. post:: Nov 22, 2022 03:00\n :language: en\n :tags: neuron2.x\n\n.. _neuron250-packages-changes:\n\nIntroducing Neuron"
},
{
"path": "about-neuron/announcements/neuron2.x/release-neuron2.4.rst",
"chars": 0,
"preview": ""
},
{
"path": "about-neuron/announcements/neuron2.x/sm-training-dlc-2.9.1.rst",
"chars": 609,
"preview": ".. post:: Apr 26, 2023 11:00\n :language: en\n :tags: sagemaker, pytorch, trn1, inf2\n\n.. _announce-dlc-sm-neuron-2.9"
},
{
"path": "about-neuron/announcements/neuron2.x/sm-training-trn1-introduce.rst",
"chars": 616,
"preview": ".. post:: Nov 03, 2022 00:01\n :language: en\n :tags: sagemaker, pytorch, trn1\n\n.. _announce-dlami-neuron-pytorch:\n\n"
},
{
"path": "about-neuron/appnotes/index.rst",
"chars": 2281,
"preview": ".. _neuron-appnotes-index:\n.. _neuron-appnotes:\n\n.. meta::\n :description: AWS Neuron SDK application notes for support"
},
{
"path": "about-neuron/appnotes/mxnet-neuron/flex-eg.rst",
"chars": 8766,
"preview": ".. _flexeg:\n\nFlexible Execution Group (FlexEG) in Neuron-MXNet\n=================================================\n\nIntrod"
},
{
"path": "about-neuron/appnotes/neuron-cc/mixed-precision.rst",
"chars": 4108,
"preview": ".. _neuron-cc-training-mixed-precision:\n\nMixed precision and performance-accuracy tuning (``neuron-cc``)\n==============="
}
]
// ... and 1436 more files (download for full content)
About this extraction
This page contains the full source code of the aws/aws-neuron-sdk GitHub repository, extracted and formatted as plain text for AI agents and large language models (LLMs). The extraction includes 1636 files (10.3 MB), approximately 2.8M tokens, and a symbol index with 1534 extracted functions, classes, methods, constants, and types. Use this with OpenClaw, Claude, ChatGPT, Cursor, Windsurf, or any other AI tool that accepts text input. You can copy the full output to your clipboard or download it as a .txt file.
Extracted by GitExtract — free GitHub repo to text converter for AI. Built by Nikandr Surkov.