Repository: nari-labs/dia Branch: main Commit: 876125e461a0 Files: 24 Total size: 141.3 KB Directory structure: gitextract_f5lpo0wj/ ├── .github/ │ └── workflows/ │ └── ci.yaml ├── .gitignore ├── .python-version ├── LICENSE ├── README.md ├── app.py ├── cli.py ├── dia/ │ ├── __init__.py │ ├── audio.py │ ├── config.py │ ├── layers.py │ ├── model.py │ └── state.py ├── docker/ │ ├── Dockerfile.cpu │ └── Dockerfile.gpu ├── example/ │ ├── benchmark.py │ ├── simple-cpu.py │ ├── simple-mac.py │ ├── simple.py │ ├── simple_batch.py │ ├── voice_clone.py │ └── voice_clone_batch.py ├── hf.py └── pyproject.toml ================================================ FILE CONTENTS ================================================ ================================================ FILE: .github/workflows/ci.yaml ================================================ name: Continuous Integration on: pull_request: branches: - main jobs: lint_and_format: runs-on: ubuntu-latest name: Lint and Format steps: - uses: actions/checkout@v4 - uses: astral-sh/ruff-action@v3 with: version: latest - name: Check Lint using Ruff run: ruff check - name: Check Format using Ruff run: ruff format --check --diff ================================================ FILE: .gitignore ================================================ # Python-generated files __pycache__/ *.py[oc] build/ dist/ wheels/ *.egg-info # Virtual environments .venv .idea/ .gradio **/*.pth **/*.safetensors **/*.mp3 !example_prompt.mp3 **/*.txt .ruff_cache .ipynb_checkpoints config.json ================================================ FILE: .python-version ================================================ 3.10 ================================================ FILE: LICENSE ================================================ Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION 1. Definitions. "License" shall mean the terms and conditions for use, reproduction, and distribution as defined by Sections 1 through 9 of this document. "Licensor" shall mean the copyright owner or entity authorized by the copyright owner that is granting the License. "Legal Entity" shall mean the union of the acting entity and all other entities that control, are controlled by, or are under common control with that entity. For the purposes of this definition, "control" means (i) the power, direct or indirect, to cause the direction or management of such entity, whether by contract or otherwise, or (ii) ownership of fifty percent (50%) or more of the outstanding shares, or (iii) beneficial ownership of such entity. "You" (or "Your") shall mean an individual or Legal Entity exercising permissions granted by this License. "Source" form shall mean the preferred form for making modifications, including but not limited to software source code, documentation source, and configuration files. "Object" form shall mean any form resulting from mechanical transformation or translation of a Source form, including but not limited to compiled object code, generated documentation, and conversions to other media types. "Work" shall mean the work of authorship, whether in Source or Object form, made available under the License, as indicated by a copyright notice that is included in or attached to the work (an example is provided in the Appendix below). "Derivative Works" shall mean any work, whether in Source or Object form, that is based on (or derived from) the Work and for which the editorial revisions, annotations, elaborations, or other modifications represent, as a whole, an original work of authorship. For the purposes of this License, Derivative Works shall not include works that remain separable from, or merely link (or bind by name) to the interfaces of, the Work and Derivative Works thereof. "Contribution" shall mean any work of authorship, including the original version of the Work and any modifications or additions to that Work or Derivative Works thereof, that is intentionally submitted to Licensor for inclusion in the Work by the copyright owner or by an individual or Legal Entity authorized to submit on behalf of the copyright owner. For the purposes of this definition, "submitted" means any form of electronic, verbal, or written communication sent to the Licensor or its representatives, including but not limited to communication on electronic mailing lists, source code control systems, and issue tracking systems that are managed by, or on behalf of, the Licensor for the purpose of discussing and improving the Work, but excluding communication that is conspicuously marked or otherwise designated in writing by the copyright owner as "Not a Contribution." "Contributor" shall mean Licensor and any individual or Legal Entity on behalf of whom a Contribution has been received by Licensor and subsequently incorporated within the Work. 2. Grant of Copyright License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to reproduce, prepare Derivative Works of, publicly display, publicly perform, sublicense, and distribute the Work and such Derivative Works in Source or Object form. 3. Grant of Patent License. Subject to the terms and conditions of this License, each Contributor hereby grants to You a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to make, have made, use, offer to sell, sell, import, and otherwise transfer the Work, where such license applies only to those patent claims licensable by such Contributor that are necessarily infringed by their Contribution(s) alone or by combination of their Contribution(s) with the Work to which such Contribution(s) was submitted. If You institute patent litigation against any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Work or a Contribution incorporated within the Work constitutes direct or contributory patent infringement, then any patent licenses granted to You under this License for that Work shall terminate as of the date such litigation is filed. 4. Redistribution. You may reproduce and distribute copies of the Work or Derivative Works thereof in any medium, with or without modifications, and in Source or Object form, provided that You meet the following conditions: (a) You must give any other recipients of the Work or Derivative Works a copy of this License; and (b) You must cause any modified files to carry prominent notices stating that You changed the files; and (c) You must retain, in the Source form of any Derivative Works that You distribute, all copyright, patent, trademark, and attribution notices from the Source form of the Work, excluding those notices that do not pertain to any part of the Derivative Works; and (d) If the Work includes a "NOTICE" text file as part of its distribution, then any Derivative Works that You distribute must include a readable copy of the attribution notices contained within such NOTICE file, excluding those notices that do not pertain to any part of the Derivative Works, in at least one of the following places: within a NOTICE text file distributed as part of the Derivative Works; within the Source form or documentation, if provided along with the Derivative Works; or, within a display generated by the Derivative Works, if and wherever such third-party notices normally appear. The contents of the NOTICE file are for informational purposes only and do not modify the License. You may add Your own attribution notices within Derivative Works that You distribute, alongside or as an addendum to the NOTICE text from the Work, provided that such additional attribution notices cannot be construed as modifying the License. You may add Your own copyright statement to Your modifications and may provide additional or different license terms and conditions for use, reproduction, or distribution of Your modifications, or for any such Derivative Works as a whole, provided Your use, reproduction, and distribution of the Work otherwise complies with the conditions stated in this License. 5. Submission of Contributions. Unless You explicitly state otherwise, any Contribution intentionally submitted for inclusion in the Work by You to the Licensor shall be under the terms and conditions of this License, without any additional terms or conditions. Notwithstanding the above, nothing herein shall supersede or modify the terms of any separate license agreement you may have executed with Licensor regarding such Contributions. 6. Trademarks. This License does not grant permission to use the trade names, trademarks, service marks, or product names of the Licensor, except as required for reasonable and customary use in describing the origin of the Work and reproducing the content of the NOTICE file. 7. Disclaimer of Warranty. Unless required by applicable law or agreed to in writing, Licensor provides the Work (and each Contributor provides its Contributions) on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied, including, without limitation, any warranties or conditions of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A PARTICULAR PURPOSE. You are solely responsible for determining the appropriateness of using or redistributing the Work and assume any risks associated with Your exercise of permissions under this License. 8. Limitation of Liability. In no event and under no legal theory, whether in tort (including negligence), contract, or otherwise, unless required by applicable law (such as deliberate and grossly negligent acts) or agreed to in writing, shall any Contributor be liable to You for damages, including any direct, indirect, special, incidental, or consequential damages of any character arising as a result of this License or out of the use or inability to use the Work (including but not limited to damages for loss of goodwill, work stoppage, computer failure or malfunction, or any and all other commercial damages or losses), even if such Contributor has been advised of the possibility of such damages. 9. Accepting Warranty or Additional Liability. While redistributing the Work or Derivative Works thereof, You may choose to offer, and charge a fee for, acceptance of support, warranty, indemnity, or other liability obligations and/or rights consistent with this License. However, in accepting such obligations, You may act only on Your own behalf and on Your sole responsibility, not on behalf of any other Contributor, and only if You agree to indemnify, defend, and hold each Contributor harmless for any liability incurred by, or claims asserted against, such Contributor by reason of your accepting any such warranty or additional liability. END OF TERMS AND CONDITIONS APPENDIX: How to apply the Apache License to your work. To apply the Apache License to your work, attach the following boilerplate notice, with the fields enclosed by brackets "[]" replaced with your own identifying information. (Don't include the brackets!) The text should be enclosed in the appropriate comment syntax for the file format. We also recommend that a file or class name and description of purpose be included on the same "printed page" as the copyright notice for easier identification within third-party archives. Copyright 2025 Nari Labs Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. ================================================ FILE: README.md ================================================

Static Badge LICENSE

Model on HuggingFace Space on HuggingFace

Dia is a 1.6B parameter text to speech model created by Nari Labs. **UPDATE 🤗(06/27)**: Dia is now available through [Hugging Face Transformers](https://github.com/huggingface/transformers)! **UPDATE 🚀(11/19)**: Dia2 is released on Github and HuggingFace [link](https://github.com/nari-labs/dia2)! Dia **directly generates highly realistic dialogue from a transcript**. You can condition the output on audio, enabling emotion and tone control. The model can also produce nonverbal communications like laughter, coughing, clearing throat, etc. To accelerate research, we are providing access to pretrained model checkpoints and inference code. The model weights are hosted on [Hugging Face](https://huggingface.co/nari-labs/Dia-1.6B-0626). The model only supports English generation at the moment. We also provide a [demo page](https://yummy-fir-7a4.notion.site/dia) comparing our model to [ElevenLabs Studio](https://elevenlabs.io/studio) and [Sesame CSM-1B](https://github.com/SesameAILabs/csm). - We have a ZeroGPU Space running! Try it now [here](https://huggingface.co/spaces/nari-labs/Dia-1.6B-0626). Thanks to the HF team for the support :) - Join our [discord server](https://discord.gg/bJq6vjRRKv) for community support and access to new features. - Play with a larger version of Dia: generate fun conversations, remix content, and share with friends. 🔮 Join the [waitlist](https://tally.so/r/meokbo) for early access. ## Generation Guidelines - Keep input text length moderate - Short input (corresponding to under 5s of audio) will sound unnatural - Very long input (corresponding to over 20s of audio) will make the speech unnaturally fast. - Use non-verbal tags sparingly, from the list in the README. Overusing or using unlisted non-verbals may cause weird artifacts. - Always begin input text with `[S1]`, and always alternate between `[S1]` and `[S2]` (i.e. `[S1]`... `[S1]`... is not good) - When using audio prompts (voice cloning), follow these instructions carefully: - Provide the transcript of the to-be cloned audio before the generation text. - Transcript must use `[S1]`, `[S2]` speaker tags correctly (i.e. single speaker: `[S1]`..., two speakers: `[S1]`... `[S2]`...) - Duration of the to-be cloned audio should be 5~10 seconds for the best results. (Keep in mind: 1 second ≈ 86 tokens) - Put `[S1]` or `[S2]` (the second-to-last speaker's tag) at the end of the audio to improve audio quality at the end ## Quickstart ### Transformers Support We now have a [Hugging Face Transformers](https://github.com/huggingface/transformers) implementation of Dia! You should install `main` branch of `transformers` to use it. See [hf.py](hf.py) for more information.
View more details Install `main` branch of `transformers` ```bash pip install git+https://github.com/huggingface/transformers.git # or install with uv uv pip install git+https://github.com/huggingface/transformers.git ``` Run `hf.py`. The file is as below. ```python from transformers import AutoProcessor, DiaForConditionalGeneration torch_device = "cuda" model_checkpoint = "nari-labs/Dia-1.6B-0626" text = [ "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." ] processor = AutoProcessor.from_pretrained(model_checkpoint) inputs = processor(text=text, padding=True, return_tensors="pt").to(torch_device) model = DiaForConditionalGeneration.from_pretrained(model_checkpoint).to(torch_device) outputs = model.generate( **inputs, max_new_tokens=3072, guidance_scale=3.0, temperature=1.8, top_p=0.90, top_k=45 ) outputs = processor.batch_decode(outputs) processor.save_audio(outputs, "example.mp3") ```
### Run with this repo
Install via pip ```bash # Clone this repository git clone https://github.com/nari-labs/dia.git cd dia # Optionally python -m venv .venv && source .venv/bin/activate # Install dia pip install -e . ``` Or you can install without cloning. ```bash # Install directly from GitHub pip install git+https://github.com/nari-labs/dia.git ``` Now, run some examples. ```bash python example/simple.py ```
Install via uv You need [uv](https://docs.astral.sh/uv/) to be installed. ```bash # Clone this repository git clone https://github.com/nari-labs/dia.git cd dia ``` Run some examples directly. ```bash uv run example/simple.py ```
Run Gradio UI ```bash python app.py # Or if you have uv installed uv run app.py ```
Run with CLI ```bash python cli.py --help # Or if you have uv installed uv run cli.py --help ```
> [!NOTE] > The model was not fine-tuned on a specific voice. Hence, you will get different voices every time you run the model. > You can keep speaker consistency by either adding an audio prompt, or fixing the seed. > [!IMPORTANT] > If you are using 5000 series GPU, you should use torch 2.8 nightly. Look at the issue [#26](https://github.com/nari-labs/dia/issues/26) for more details. ## Features - Generate dialogue via `[S1]` and `[S2]` tag - Generate non-verbal like `(laughs)`, `(coughs)`, etc. - Below verbal tags will be recognized, but might result in unexpected output. - `(laughs), (clears throat), (sighs), (gasps), (coughs), (singing), (sings), (mumbles), (beep), (groans), (sniffs), (claps), (screams), (inhales), (exhales), (applause), (burps), (humming), (sneezes), (chuckle), (whistles)` - Voice cloning. See [`example/voice_clone.py`](example/voice_clone.py) for more information. - In the Hugging Face space, you can upload the audio you want to clone and place its transcript before your script. Make sure the transcript follows the required format. The model will then output only the content of your script. ## 💻 Hardware and Inference Speed Dia has been tested on only GPUs (pytorch 2.0+, CUDA 12.6). CPU support is to be added soon. The initial run will take longer as the Descript Audio Codec also needs to be downloaded. These are the speed we benchmarked in RTX 4090. | precision | realtime factor w/ compile | realtime factor w/o compile | VRAM | |:-:|:-:|:-:|:-:| | `bfloat16` | x2.1 | x1.5 | ~4.4GB | | `float16` | x2.2 | x1.3 | ~4.4GB | | `float32` | x1 | x0.9 | ~7.9GB | We will be adding a quantized version in the future. If you don't have hardware available or if you want to play with bigger versions of our models, join the waitlist [here](https://tally.so/r/meokbo). ## 🪪 License This project is licensed under the Apache License 2.0 - see the [LICENSE](LICENSE) file for details. ## ⚠️ Disclaimer This project offers a high-fidelity speech generation model intended for research and educational use. The following uses are **strictly forbidden**: - **Identity Misuse**: Do not produce audio resembling real individuals without permission. - **Deceptive Content**: Do not use this model to generate misleading content (e.g. fake news) - **Illegal or Malicious Use**: Do not use this model for activities that are illegal or intended to cause harm. By using this model, you agree to uphold relevant legal standards and ethical responsibilities. We **are not responsible** for any misuse and firmly oppose any unethical usage of this technology. ## 🔭 TODO / Future Work - Docker support for ARM architecture and MacOS. - Optimize inference speed. - Add quantization for memory efficiency. ## 🤝 Contributing We are a tiny team of 1 full-time and 1 part-time research-engineers. We are extra-welcome to any contributions! Join our [Discord Server](https://discord.gg/bJq6vjRRKv) for discussions. ## 🤗 Acknowledgements - We thank the [Google TPU Research Cloud program](https://sites.research.google/trc/about/) for providing computation resources. - Our work was heavily inspired by [SoundStorm](https://arxiv.org/abs/2305.09636), [Parakeet](https://jordandarefsky.com/blog/2024/parakeet/), and [Descript Audio Codec](https://github.com/descriptinc/descript-audio-codec). - Hugging Face for providing the ZeroGPU Grant. - "Nari" is a pure Korean word for lily. - We thank Jason Y. for providing help with data filtering. ## ⭐ Star History Star History Chart ================================================ FILE: app.py ================================================ import argparse import contextlib import io import random import tempfile import time from pathlib import Path from typing import Optional, Tuple import gradio as gr import numpy as np import soundfile as sf import torch from dia.model import Dia # --- Global Setup --- parser = argparse.ArgumentParser(description="Gradio interface for Nari TTS") parser.add_argument("--device", type=str, default=None, help="Force device (e.g., 'cuda', 'mps', 'cpu')") parser.add_argument("--share", action="store_true", help="Enable Gradio sharing") args = parser.parse_args() # Determine device if args.device: device = torch.device(args.device) elif torch.cuda.is_available(): device = torch.device("cuda") # Simplified MPS check for broader compatibility elif hasattr(torch.backends, "mps") and torch.backends.mps.is_available(): # Basic check is usually sufficient, detailed check can be problematic device = torch.device("mps") else: device = torch.device("cpu") print(f"Using device: {device}") # Load Nari model and config print("Loading Nari model...") try: dtype_map = { "cpu": "float32", "mps": "float32", # Apple M series – better with float32 "cuda": "float16", # NVIDIA – better with float16 } dtype = dtype_map.get(device.type, "float16") print(f"Using device: {device}, attempting to load model with {dtype}") model = Dia.from_pretrained("nari-labs/Dia-1.6B-0626", compute_dtype=dtype, device=device) except Exception as e: print(f"Error loading Nari model: {e}") raise def set_seed(seed: int): """Sets the random seed for reproducibility.""" random.seed(seed) np.random.seed(seed) torch.manual_seed(seed) if torch.cuda.is_available(): torch.cuda.manual_seed(seed) torch.cuda.manual_seed_all(seed) torch.backends.cudnn.deterministic = True torch.backends.cudnn.benchmark = False def run_inference( text_input: str, audio_prompt_text_input: str, audio_prompt_input: Optional[Tuple[int, np.ndarray]], max_new_tokens: int, cfg_scale: float, temperature: float, top_p: float, cfg_filter_top_k: int, speed_factor: float, seed: Optional[int] = None, ): """ Runs Nari inference using the globally loaded model and provided inputs. Uses temporary files for text and audio prompt compatibility with inference.generate. """ global model, device # Access global model, config, device console_output_buffer = io.StringIO() with contextlib.redirect_stdout(console_output_buffer): # Prepend transcript text if audio_prompt provided if audio_prompt_input and audio_prompt_text_input and not audio_prompt_text_input.isspace(): text_input = audio_prompt_text_input + "\n" + text_input text_input = text_input.strip() if audio_prompt_input and (not audio_prompt_text_input or audio_prompt_text_input.isspace()): raise gr.Error("Audio Prompt Text input cannot be empty.") if not text_input or text_input.isspace(): raise gr.Error("Text input cannot be empty.") # Preprocess Audio temp_txt_file_path = None temp_audio_prompt_path = None output_audio = (44100, np.zeros(1, dtype=np.float32)) try: prompt_path_for_generate = None if audio_prompt_input is not None: sr, audio_data = audio_prompt_input # Check if audio_data is valid if audio_data is None or audio_data.size == 0 or audio_data.max() == 0: # Check for silence/empty gr.Warning("Audio prompt seems empty or silent, ignoring prompt.") else: # Save prompt audio to a temporary WAV file with tempfile.NamedTemporaryFile(mode="wb", suffix=".wav", delete=False) as f_audio: temp_audio_prompt_path = f_audio.name # Store path for cleanup # Basic audio preprocessing for consistency # Convert to float32 in [-1, 1] range if integer type if np.issubdtype(audio_data.dtype, np.integer): max_val = np.iinfo(audio_data.dtype).max audio_data = audio_data.astype(np.float32) / max_val elif not np.issubdtype(audio_data.dtype, np.floating): gr.Warning(f"Unsupported audio prompt dtype {audio_data.dtype}, attempting conversion.") # Attempt conversion, might fail for complex types try: audio_data = audio_data.astype(np.float32) except Exception as conv_e: raise gr.Error(f"Failed to convert audio prompt to float32: {conv_e}") # Ensure mono (average channels if stereo) if audio_data.ndim > 1: if audio_data.shape[0] == 2: # Assume (2, N) audio_data = np.mean(audio_data, axis=0) elif audio_data.shape[1] == 2: # Assume (N, 2) audio_data = np.mean(audio_data, axis=1) else: gr.Warning( f"Audio prompt has unexpected shape {audio_data.shape}, taking first channel/axis." ) audio_data = ( audio_data[0] if audio_data.shape[0] < audio_data.shape[1] else audio_data[:, 0] ) audio_data = np.ascontiguousarray(audio_data) # Ensure contiguous after slicing/mean # Write using soundfile try: sf.write( temp_audio_prompt_path, audio_data, sr, subtype="FLOAT" ) # Explicitly use FLOAT subtype prompt_path_for_generate = temp_audio_prompt_path print(f"Created temporary audio prompt file: {temp_audio_prompt_path} (orig sr: {sr})") except Exception as write_e: print(f"Error writing temporary audio file: {write_e}") raise gr.Error(f"Failed to save audio prompt: {write_e}") # Set and Display Generation Seed if seed is None or seed < 0: seed = random.randint(0, 2**32 - 1) print(f"\nNo seed provided, generated random seed: {seed}\n") else: print(f"\nUsing user-selected seed: {seed}\n") set_seed(seed) # Run Generation print(f'Generating speech: \n"{text_input}"\n') start_time = time.time() # Use torch.inference_mode() context manager for the generation call with torch.inference_mode(): output_audio_np = model.generate( text_input, max_tokens=max_new_tokens, cfg_scale=cfg_scale, temperature=temperature, top_p=top_p, cfg_filter_top_k=cfg_filter_top_k, # Pass the value here use_torch_compile=False, # Keep False for Gradio stability audio_prompt=prompt_path_for_generate, verbose=True, ) end_time = time.time() print(f"Generation finished in {end_time - start_time:.2f} seconds.\n") # 4. Convert Codes to Audio if output_audio_np is not None: # Get sample rate from the loaded DAC model output_sr = 44100 # --- Slow down audio --- original_len = len(output_audio_np) # Ensure speed_factor is positive and not excessively small/large to avoid issues speed_factor = max(0.1, min(speed_factor, 5.0)) target_len = int(original_len / speed_factor) # Target length based on speed_factor if target_len != original_len and target_len > 0: # Only interpolate if length changes and is valid x_original = np.arange(original_len) x_resampled = np.linspace(0, original_len - 1, target_len) resampled_audio_np = np.interp(x_resampled, x_original, output_audio_np) output_audio = ( output_sr, resampled_audio_np.astype(np.float32), ) # Use resampled audio print( f"Resampled audio from {original_len} to {target_len} samples for {speed_factor:.2f}x speed." ) else: output_audio = ( output_sr, output_audio_np, ) # Keep original if calculation fails or no change print(f"Skipping audio speed adjustment (factor: {speed_factor:.2f}).") # --- End slowdown --- print(f"Audio conversion successful. Final shape: {output_audio[1].shape}, Sample Rate: {output_sr}") # Explicitly convert to int16 to prevent Gradio warning if output_audio[1].dtype == np.float32 or output_audio[1].dtype == np.float64: audio_for_gradio = np.clip(output_audio[1], -1.0, 1.0) audio_for_gradio = (audio_for_gradio * 32767).astype(np.int16) output_audio = (output_sr, audio_for_gradio) print("Converted audio to int16 for Gradio output.") else: print("\nGeneration finished, but no valid tokens were produced.") # Return default silence gr.Warning("Generation produced no output.") except Exception as e: print(f"Error during inference: {e}") import traceback traceback.print_exc() # Re-raise as Gradio error to display nicely in the UI raise gr.Error(f"Inference failed: {e}") finally: # Cleanup Temporary Files defensively if temp_txt_file_path and Path(temp_txt_file_path).exists(): try: Path(temp_txt_file_path).unlink() print(f"Deleted temporary text file: {temp_txt_file_path}") except OSError as e: print(f"Warning: Error deleting temporary text file {temp_txt_file_path}: {e}") if temp_audio_prompt_path and Path(temp_audio_prompt_path).exists(): try: Path(temp_audio_prompt_path).unlink() print(f"Deleted temporary audio prompt file: {temp_audio_prompt_path}") except OSError as e: print(f"Warning: Error deleting temporary audio prompt file {temp_audio_prompt_path}: {e}") # After generation, capture the printed output console_output = console_output_buffer.getvalue() return output_audio, seed, console_output # --- Create Gradio Interface --- css = """ #col-container {max-width: 90%; margin-left: auto; margin-right: auto;} """ # Attempt to load default text from example.txt default_text = "[S1] Dia is an open weights text to dialogue model. \n[S2] You get full control over scripts and voices. \n[S1] Wow. Amazing. (laughs) \n[S2] Try it now on Git hub or Hugging Face." example_txt_path = Path("./example.txt") if example_txt_path.exists(): try: default_text = example_txt_path.read_text(encoding="utf-8").strip() if not default_text: # Handle empty example file default_text = "Example text file was empty." except Exception as e: print(f"Warning: Could not read example.txt: {e}") # Build Gradio UI with gr.Blocks(css=css, theme="gradio/dark") as demo: gr.Markdown("# Nari Text-to-Speech Synthesis") with gr.Row(equal_height=False): with gr.Column(scale=1): with gr.Accordion("Audio Reference Prompt (Optional)", open=False): audio_prompt_input = gr.Audio( label="Audio Prompt (Optional)", show_label=True, sources=["upload", "microphone"], type="numpy", ) audio_prompt_text_input = gr.Textbox( label="Transcript of Audio Prompt (Required if using Audio Prompt)", placeholder="Enter text here...", value="", lines=5, # Increased lines ) text_input = gr.Textbox( label="Text To Generate", placeholder="Enter text here...", value=default_text, lines=5, # Increased lines ) with gr.Accordion("Generation Parameters", open=False): max_new_tokens = gr.Slider( label="Max New Tokens (Audio Length)", minimum=860, maximum=3072, value=model.config.decoder_config.max_position_embeddings, # Use config default if available, else fallback step=50, info="Controls the maximum length of the generated audio (more tokens = longer audio).", ) cfg_scale = gr.Slider( label="CFG Scale (Guidance Strength)", minimum=1.0, maximum=5.0, value=3.0, # Default from inference.py step=0.1, info="Higher values increase adherence to the text prompt.", ) temperature = gr.Slider( label="Temperature (Randomness)", minimum=1.0, maximum=2.5, value=1.8, # Default from inference.py step=0.05, info="Lower values make the output more deterministic, higher values increase randomness.", ) top_p = gr.Slider( label="Top P (Nucleus Sampling)", minimum=0.70, maximum=1.0, value=0.95, # Default from inference.py step=0.01, info="Filters vocabulary to the most likely tokens cumulatively reaching probability P.", ) cfg_filter_top_k = gr.Slider( label="CFG Filter Top K", minimum=15, maximum=100, value=45, step=1, info="Top k filter for CFG guidance.", ) speed_factor_slider = gr.Slider( label="Speed Factor", minimum=0.8, maximum=1.0, value=1.0, step=0.02, info="Adjusts the speed of the generated audio (1.0 = original speed).", ) seed_input = gr.Number( label="Generation Seed (Optional)", value=-1, precision=0, # No decimal points step=1, interactive=True, info="Set a generation seed for reproducible outputs. Leave empty or -1 for random seed.", ) run_button = gr.Button("Generate Audio", variant="primary") with gr.Column(scale=1): audio_output = gr.Audio( label="Generated Audio", type="numpy", autoplay=False, ) seed_output = gr.Textbox(label="Generation Seed", interactive=False) console_output = gr.Textbox(label="Console Output Log", lines=10, interactive=False) # Link button click to function run_button.click( fn=run_inference, inputs=[ text_input, audio_prompt_text_input, audio_prompt_input, max_new_tokens, cfg_scale, temperature, top_p, cfg_filter_top_k, speed_factor_slider, seed_input, ], outputs=[ audio_output, seed_output, console_output, ], # Add status_output here if using it api_name="generate_audio", ) # Add examples (ensure the prompt path is correct or remove it if example file doesn't exist) example_prompt_path = "./example_prompt.mp3" # Adjust if needed examples_list = [ [ "[S1] Oh fire! Oh my goodness! What's the procedure? What to we do people? The smoke could be coming through an air duct! \n[S2] Oh my god! Okay.. it's happening. Everybody stay calm! \n[S1] What's the procedure... \n[S2] Everybody stay fucking calm!!!... Everybody fucking calm down!!!!! \n[S1] No! No! If you touch the handle, if its hot there might be a fire down the hallway! ", None, 3072, 3.0, 1.8, 0.95, 45, 1.0, ], [ "[S1] Open weights text to dialogue model. \n[S2] You get full control over scripts and voices. \n[S1] I'm biased, but I think we clearly won. \n[S2] Hard to disagree. (laughs) \n[S1] Thanks for listening to this demo. \n[S2] Try it now on Git hub and Hugging Face. \n[S1] If you liked our model, please give us a star and share to your friends. \n[S2] This was Nari Labs.", example_prompt_path if Path(example_prompt_path).exists() else None, 3072, 3.0, 1.8, 0.95, 45, 1.0, ], ] if examples_list: gr.Examples( examples=examples_list, inputs=[ text_input, audio_prompt_input, max_new_tokens, cfg_scale, temperature, top_p, cfg_filter_top_k, speed_factor_slider, seed_input, ], outputs=[audio_output], fn=run_inference, cache_examples=False, label="Examples (Click to Run)", ) else: gr.Markdown("_(No examples configured or example prompt file missing)_") # --- Launch the App --- if __name__ == "__main__": print("Launching Gradio interface...") # set `GRADIO_SERVER_NAME`, `GRADIO_SERVER_PORT` env vars to override default values # use `GRADIO_SERVER_NAME=0.0.0.0` for Docker demo.launch(share=args.share) ================================================ FILE: cli.py ================================================ import argparse import os import random import numpy as np import soundfile as sf import torch from dia.model import Dia def set_seed(seed: int): """Sets the random seed for reproducibility.""" random.seed(seed) np.random.seed(seed) torch.manual_seed(seed) if torch.cuda.is_available(): torch.cuda.manual_seed(seed) torch.cuda.manual_seed_all(seed) # Ensure deterministic behavior for cuDNN (if used) torch.backends.cudnn.deterministic = True torch.backends.cudnn.benchmark = False def main(): parser = argparse.ArgumentParser(description="Generate audio using the Dia model.") parser.add_argument("text", type=str, help="Input text for speech generation.") parser.add_argument( "--output", type=str, required=True, help="Path to save the generated audio file (e.g., output.wav)." ) parser.add_argument( "--repo-id", type=str, default="nari-labs/Dia-1.6B-0626", help="Hugging Face repository ID (e.g., nari-labs/Dia-1.6B-0626).", ) parser.add_argument( "--local-paths", action="store_true", help="Load model from local config and checkpoint files." ) parser.add_argument( "--config", type=str, help="Path to local config.json file (required if --local-paths is set)." ) parser.add_argument( "--checkpoint", type=str, help="Path to local model checkpoint .pth file (required if --local-paths is set)." ) parser.add_argument( "--audio-prompt", type=str, default=None, help="Path to an optional audio prompt WAV file for voice cloning." ) gen_group = parser.add_argument_group("Generation Parameters") gen_group.add_argument( "--max-tokens", type=int, default=None, help="Maximum number of audio tokens to generate (defaults to config value).", ) gen_group.add_argument( "--cfg-scale", type=float, default=3.0, help="Classifier-Free Guidance scale (default: 3.0)." ) gen_group.add_argument( "--temperature", type=float, default=1.3, help="Sampling temperature (higher is more random, default: 0.7)." ) gen_group.add_argument("--top-p", type=float, default=0.95, help="Nucleus sampling probability (default: 0.95).") infra_group = parser.add_argument_group("Infrastructure") infra_group.add_argument("--seed", type=int, default=None, help="Random seed for reproducibility.") infra_group.add_argument( "--device", type=str, default="cuda" if torch.cuda.is_available() else "cpu", help="Device to run inference on (e.g., 'cuda', 'cpu', default: auto).", ) args = parser.parse_args() # Validation for local paths if args.local_paths: if not args.config: parser.error("--config is required when --local-paths is set.") if not args.checkpoint: parser.error("--checkpoint is required when --local-paths is set.") if not os.path.exists(args.config): parser.error(f"Config file not found: {args.config}") if not os.path.exists(args.checkpoint): parser.error(f"Checkpoint file not found: {args.checkpoint}") # Set seed if provided if args.seed is not None: set_seed(args.seed) print(f"Using user-selected seed: {args.seed}") # Determine device device = torch.device(args.device) print(f"Using device: {device}") # Load model print("Loading model...") if args.local_paths: print(f"Loading from local paths: config='{args.config}', checkpoint='{args.checkpoint}'") try: model = Dia.from_local(args.config, args.checkpoint, device=device) except Exception as e: print(f"Error loading local model: {e}") exit(1) else: print(f"Loading from Hugging Face Hub: repo_id='{args.repo_id}'") try: model = Dia.from_pretrained(args.repo_id, device=device) except Exception as e: print(f"Error loading model from Hub: {e}") exit(1) print("Model loaded.") # Generate audio print("Generating audio...") try: sample_rate = 44100 # Default assumption output_audio = model.generate( text=args.text, audio_prompt=args.audio_prompt, max_tokens=args.max_tokens, cfg_scale=args.cfg_scale, temperature=args.temperature, top_p=args.top_p, ) print("Audio generation complete.") print(f"Saving audio to {args.output}...") os.makedirs(os.path.dirname(args.output) or ".", exist_ok=True) sf.write(args.output, output_audio, sample_rate) print(f"Audio successfully saved to {args.output}") except Exception as e: print(f"Error during audio generation or saving: {e}") exit(1) if __name__ == "__main__": main() ================================================ FILE: dia/__init__.py ================================================ from .model import Dia __all__ = [ "Dia", ] ================================================ FILE: dia/audio.py ================================================ import typing as tp import torch def build_delay_indices(B: int, T: int, C: int, delay_pattern: tp.List[int]) -> tp.Tuple[torch.Tensor, torch.Tensor]: """ Precompute (t_idx_BxTxC, indices_BTCx3) so that out[t, c] = in[t - delay[c], c]. Negative t_idx => BOS; t_idx >= T => PAD. """ delay_arr = torch.tensor(delay_pattern, dtype=torch.int32) t_idx_BxT = torch.broadcast_to( torch.arange(T, dtype=torch.int32)[None, :], [B, T], ) t_idx_BxTx1 = t_idx_BxT[..., None] t_idx_BxTxC = t_idx_BxTx1 - delay_arr.view(1, 1, C) b_idx_BxTxC = torch.broadcast_to( torch.arange(B, dtype=torch.int32).view(B, 1, 1), [B, T, C], ) c_idx_BxTxC = torch.broadcast_to( torch.arange(C, dtype=torch.int32).view(1, 1, C), [B, T, C], ) # We must clamp time indices to [0..T-1] so gather_nd equivalent won't fail t_clamped_BxTxC = torch.clamp(t_idx_BxTxC, 0, T - 1) indices_BTCx3 = torch.stack( [ b_idx_BxTxC.reshape(-1), t_clamped_BxTxC.reshape(-1), c_idx_BxTxC.reshape(-1), ], dim=1, ).long() # Ensure indices are long type for indexing return t_idx_BxTxC, indices_BTCx3 def apply_audio_delay( audio_BxTxC: torch.Tensor, pad_value: int, bos_value: int, precomp: tp.Tuple[torch.Tensor, torch.Tensor], ) -> torch.Tensor: """ Applies the delay pattern to batched audio tokens using precomputed indices, inserting BOS where t_idx < 0 and PAD where t_idx >= T. Args: audio_BxTxC: [B, T, C] int16 audio tokens (or int32/float) pad_value: the padding token bos_value: the BOS token precomp: (t_idx_BxTxC, indices_BTCx3) from build_delay_indices Returns: result_BxTxC: [B, T, C] delayed audio tokens """ device = audio_BxTxC.device # Get device from input tensor t_idx_BxTxC, indices_BTCx3 = precomp t_idx_BxTxC = t_idx_BxTxC.to(device) # Move precomputed indices to device indices_BTCx3 = indices_BTCx3.to(device) # Equivalent of tf.gather_nd using advanced indexing # Ensure indices are long type if not already (build_delay_indices should handle this) gathered_flat = audio_BxTxC[indices_BTCx3[:, 0], indices_BTCx3[:, 1], indices_BTCx3[:, 2]] gathered_BxTxC = gathered_flat.view(audio_BxTxC.shape) # Create masks on the correct device mask_bos = t_idx_BxTxC < 0 # => place bos_value mask_pad = t_idx_BxTxC >= audio_BxTxC.shape[1] # => place pad_value # Create scalar tensors on the correct device bos_tensor = torch.tensor(bos_value, dtype=audio_BxTxC.dtype, device=device) pad_tensor = torch.tensor(pad_value, dtype=audio_BxTxC.dtype, device=device) # If mask_bos, BOS; else if mask_pad, PAD; else original gather # All tensors should now be on the same device result_BxTxC = torch.where(mask_bos, bos_tensor, torch.where(mask_pad, pad_tensor, gathered_BxTxC)) return result_BxTxC def build_revert_indices(B: int, T: int, C: int, delay_pattern: tp.List[int]) -> tp.Tuple[torch.Tensor, torch.Tensor]: """ Precompute indices for the revert operation using PyTorch. Returns: A tuple (t_idx_BxTxC, indices_BTCx3) where: - t_idx_BxTxC is a tensor of shape [B, T, C] computed as time indices plus the delay. - indices_BTCx3 is a tensor of shape [B*T*C, 3] used for gathering, computed from: batch indices, clamped time indices, and channel indices. """ # Use default device unless specified otherwise; assumes inputs might define device later device = None # Or determine dynamically if needed, e.g., from a model parameter delay_arr = torch.tensor(delay_pattern, dtype=torch.int32, device=device) t_idx_BT1 = torch.broadcast_to(torch.arange(T, device=device).unsqueeze(0), [B, T]) t_idx_BT1 = t_idx_BT1.unsqueeze(-1) t_idx_BxTxC = torch.minimum( t_idx_BT1 + delay_arr.view(1, 1, C), torch.tensor(T - 1, device=device), ) b_idx_BxTxC = torch.broadcast_to(torch.arange(B, device=device).view(B, 1, 1), [B, T, C]) c_idx_BxTxC = torch.broadcast_to(torch.arange(C, device=device).view(1, 1, C), [B, T, C]) indices_BTCx3 = torch.stack( [ b_idx_BxTxC.reshape(-1), t_idx_BxTxC.reshape(-1), c_idx_BxTxC.reshape(-1), ], axis=1, ).long() # Ensure indices are long type return t_idx_BxTxC, indices_BTCx3 def revert_audio_delay( audio_BxTxC: torch.Tensor, pad_value: int, precomp: tp.Tuple[torch.Tensor, torch.Tensor], T: int, ) -> torch.Tensor: """ Reverts a delay pattern from batched audio tokens using precomputed indices (PyTorch version). Args: audio_BxTxC: Input delayed audio tensor pad_value: Padding value for out-of-bounds indices precomp: Precomputed revert indices tuple containing: - t_idx_BxTxC: Time offset indices tensor - indices_BTCx3: Gather indices tensor for original audio T: Original sequence length before padding Returns: Reverted audio tensor with same shape as input """ t_idx_BxTxC, indices_BTCx3 = precomp device = audio_BxTxC.device # Get device from input tensor # Move precomputed indices to the same device as audio_BxTxC if they aren't already t_idx_BxTxC = t_idx_BxTxC.to(device) indices_BTCx3 = indices_BTCx3.to(device) # Using PyTorch advanced indexing (equivalent to tf.gather_nd or np equivalent) gathered_flat = audio_BxTxC[indices_BTCx3[:, 0], indices_BTCx3[:, 1], indices_BTCx3[:, 2]] gathered_BxTxC = gathered_flat.view(audio_BxTxC.size()) # Use .size() for robust reshaping # Create pad_tensor on the correct device pad_tensor = torch.tensor(pad_value, dtype=audio_BxTxC.dtype, device=device) # Create T tensor on the correct device for comparison T_tensor = torch.tensor(T, device=device) result_BxTxC = torch.where(t_idx_BxTxC >= T_tensor, pad_tensor, gathered_BxTxC) # Changed np.where to torch.where return result_BxTxC ================================================ FILE: dia/config.py ================================================ """Configuration management module for the Dia model. This module provides comprehensive configuration management for the Dia model, utilizing Pydantic for validation. It defines configurations for data processing, model architecture (encoder and decoder), and training settings. Key components: - DataConfig: Parameters for data loading and preprocessing. - EncoderConfig: Architecture details for the encoder module. - DecoderConfig: Architecture details for the decoder module. - ModelConfig: Combined model architecture settings. - TrainingConfig: Training hyperparameters and settings. - DiaConfig: Master configuration combining all components. """ import os from pydantic import BaseModel, Field class EncoderConfig(BaseModel, frozen=True): """Configuration for the encoder component of the Dia model. Attributes: model_type: Type of the model, defaults to "dia_encoder". hidden_size: Size of the encoder layers, defaults to 1024. intermediate_size: Size of the "intermediate" (i.e., feed-forward) layer in the encoder, defaults to 4096. num_hidden_layers: Number of hidden layers in the encoder, defaults to 12. num_attention_heads: Number of attention heads in the encoder, defaults to 16. num_key_value_heads: Number of key-value heads in the encoder, defaults to 16. head_dim: Dimension of each attention head, defaults to 128. hidden_act: Activation function in the encoder, defaults to "silu". max_position_embeddings: Maximum number of position embeddings, defaults to 1024. initializer_range: Range for initializing weights, defaults to 0.02. norm_eps: Epsilon value for normalization layers, defaults to 1e-5. rope_theta: Theta value for RoPE, defaults to 10000.0. rope_scaling: Optional scaling factor for RoPE. vocab_size: Vocabulary size, defaults to 256. """ head_dim: int = Field(default=128, gt=0) hidden_act: str = Field(default="silu") hidden_size: int = Field(default=1024, gt=0) initializer_range: float = Field(default=0.02) intermediate_size: int = Field(default=4096, gt=0) max_position_embeddings: int = Field(default=1024, gt=0) model_type: str = Field(default="dia_encoder") norm_eps: float = Field(default=1e-5) num_attention_heads: int = Field(default=16, gt=0) num_hidden_layers: int = Field(default=12, gt=0) num_key_value_heads: int = Field(default=16, gt=0) rope_scaling: float | None = Field(default=None) rope_theta: float = Field(default=10000.0) vocab_size: int = Field(default=256, gt=0) class DecoderConfig(BaseModel, frozen=True): """Configuration for the decoder component of the Dia model. Attributes: model_type: Type of the model, defaults to "dia_decoder". hidden_size: Size of the decoder layers, defaults to 2048. intermediate_size: Size of the "intermediate" (i.e., feed-forward) layer in the decoder, defaults to 8192. num_hidden_layers: Number of hidden layers in the decoder, defaults to 18. num_attention_heads: Number of attention heads in the decoder, defaults to 16. num_key_value_heads: Number of key-value heads in the decoder, defaults to 4. head_dim: Dimension of each attention head, defaults to 128. cross_hidden_size: Size of the cross-attention layers, defaults to 1024. cross_num_attention_heads: Number of attention heads in the cross-attention mechanism, defaults to 16. cross_num_key_value_heads: Number of key-value heads in the cross-attention mechanism, defaults to 16. cross_head_dim: Dimension of each cross-attention head, defaults to 128. hidden_act: Activation function in the decoder, defaults to "silu". max_position_embeddings: Maximum number of position embeddings in the decoder, defaults to 3072. initializer_range: Range for initializing weights in the decoder, defaults to 0.02. norm_eps: Epsilon value for normalization layers in the decoder, defaults to 1e-5. rope_theta: Theta value for RoPE in the decoder, defaults to 10000.0. rope_scaling: Optional scaling factor for RoPE in the decoder. vocab_size: Vocabulary size for the decoder, defaults to 1028. num_channels: Number of channels in the decoder, defaults to 9. """ cross_head_dim: int = Field(default=128, gt=0) cross_hidden_size: int = Field(default=1024, gt=0) cross_num_attention_heads: int = Field(default=16, gt=0) cross_num_key_value_heads: int = Field(default=16, gt=0) head_dim: int = Field(default=128, gt=0) hidden_act: str = Field(default="silu") hidden_size: int = Field(default=2048, gt=0) initializer_range: float = Field(default=0.02) intermediate_size: int = Field(default=8192, gt=0) max_position_embeddings: int = Field(default=3072, gt=0) model_type: str = Field(default="dia_decoder") norm_eps: float = Field(default=1e-5) num_attention_heads: int = Field(default=16, gt=0) num_channels: int = Field(default=9, gt=0) num_hidden_layers: int = Field(default=18, gt=0) num_key_value_heads: int = Field(default=4, gt=0) rope_scaling: float | None = Field(default=None) rope_theta: float = Field(default=10000.0) vocab_size: int = Field(default=1028, gt=0) class DiaConfig(BaseModel, frozen=True): """Main configuration container for the Dia model architecture. Attributes: model_type: Type of the model, defaults to "dia". is_encoder_decoder: Flag indicating if the model is an encoder-decoder type, defaults to True. encoder: Configuration for the encoder component. decoder: Configuration for the decoder component. src_vocab_size: Size of the source (text) vocabulary. tgt_vocab_size: Size of the target (audio code) vocabulary. initializer_range: Range for initializing weights, defaults to 0.02. norm_eps: Epsilon value for normalization layers, defaults to 1e-5. torch_dtype: Data type for model weights in PyTorch, defaults to "float32". bos_token_id: Beginning-of-sequence token ID, defaults to 1026. eos_token_id: End-of-sequence token ID, defaults to 1024. pad_token_id: Padding token ID, defaults to 1025. rope_theta: Theta value for RoPE, defaults to 10000.0. rope_scaling: Optional scaling factor for RoPE. transformers_version: Version of the transformers library, defaults to "4.53.0.dev0". architectures: List of model architectures, defaults to ["DiaForConditionalGeneration"]. delay_pattern: List of delay values for each audio channel, defaults to [0,8,9,10,11,12,13,14,15]. """ architectures: list[str] = Field(default_factory=lambda: ["DiaForConditionalGeneration"]) bos_token_id: int = Field(default=1026) decoder_config: DecoderConfig delay_pattern: list[int] = Field(default_factory=lambda: [0, 8, 9, 10, 11, 12, 13, 14, 15]) encoder_config: EncoderConfig eos_token_id: int = Field(default=1024) initializer_range: float = Field(default=0.02) is_encoder_decoder: bool = Field(default=True) model_type: str = Field(default="dia") norm_eps: float = Field(default=1e-5) pad_token_id: int = Field(default=1025) torch_dtype: str = Field(default="float32") transformers_version: str = Field(default="4.53.0.dev0") def save(self, path: str) -> None: """Save the current configuration instance to a JSON file. Ensures the parent directory exists and the file has a .json extension. Args: path: The target file path to save the configuration. Raises: ValueError: If the path is not a file with a .json extension. """ os.makedirs(os.path.dirname(path), exist_ok=True) config_json = self.model_dump_json(indent=2) with open(path, "w") as f: f.write(config_json) @classmethod def load(cls, path: str) -> "DiaConfig | None": """Load and validate a Dia configuration from a JSON file. Args: path: The path to the configuration file. Returns: A validated DiaConfig instance if the file exists and is valid, otherwise None if the file is not found. Raises: ValueError: If the path does not point to an existing .json file. pydantic.ValidationError: If the JSON content fails validation against the DiaConfig schema. """ try: with open(path, "r") as f: content = f.read() return cls.model_validate_json(content) except FileNotFoundError: return None ================================================ FILE: dia/layers.py ================================================ import torch import torch.nn as nn import torch.nn.functional as F from huggingface_hub import PyTorchModelHubMixin from torch import Tensor from torch.nn import RMSNorm from .config import DecoderConfig, DiaConfig, EncoderConfig from .state import DecoderInferenceState, EncoderInferenceState, KVCache def _normalize_axes(axes: tuple[int, ...], ndim: int) -> tuple[int, ...]: return tuple(ax if ax >= 0 else ndim + ax for ax in axes) class DenseGeneral(nn.Module): """ PyTorch equivalent of flax.linen.DenseGeneral with shapes defined at init. Stores weights (`kernel`) in the same layout as Jax and uses torch.tensordot for the generalized matrix multiplication. Weight/bias shapes are calculated and parameters created during initialization based on config. `load_weights` validates shapes and copies data. Attributes: axis (Tuple[int, ...]): Input axis or axes to contract. in_shapes (Tuple[int, ...]): Sizes of the input dimensions specified by `axis`. out_features (Tuple[int, ...]): Shape of the output features (non-contracted dims). use_bias (bool): Whether to add a bias term. weight (nn.Parameter): The kernel parameter. bias (Optional[nn.Parameter]): The bias parameter (if use_bias=True). """ def __init__( self, in_shapes: tuple[int, ...], out_features: tuple[int, ...], axis: tuple[int, ...] = (-1,), weight_dtype: torch.dtype | None = None, device: torch.device | None = None, ): super().__init__() self.in_shapes = in_shapes self.out_features = out_features self.axis = axis self.kernel_shape = self.in_shapes + self.out_features factory_kwargs = {"device": device, "dtype": weight_dtype} self.weight = nn.Parameter(torch.empty(self.kernel_shape, **factory_kwargs)) def forward(self, inputs: Tensor) -> Tensor: norm_axis = _normalize_axes(self.axis, inputs.ndim) kernel_contract_axes = tuple(range(len(norm_axis))) output = torch.tensordot( inputs.to(self.weight.dtype), self.weight, dims=(norm_axis, kernel_contract_axes), ).to(inputs.dtype) return output class MlpBlock(nn.Module): """MLP block using DenseGeneral.""" def __init__(self, embed_dim: int, intermediate_dim: int, compute_dtype: torch.dtype): super().__init__() self.dtype = compute_dtype self.wi_fused = DenseGeneral( in_shapes=(embed_dim,), out_features=(2, intermediate_dim), axis=(-1,), weight_dtype=compute_dtype, ) self.wo = DenseGeneral( in_shapes=(intermediate_dim,), out_features=(embed_dim,), axis=(-1,), weight_dtype=compute_dtype, ) def forward(self, x: torch.Tensor) -> torch.Tensor: """Forward pass.""" fused_x = self.wi_fused(x) gate = fused_x[..., 0, :] up = fused_x[..., 1, :] hidden = torch.mul(F.silu(gate), up).to(self.dtype) output = self.wo(hidden) return output class RotaryEmbedding(nn.Module): """Rotary Position Embedding (RoPE) implementation in PyTorch.""" def __init__( self, embedding_dims: int, min_timescale: float = 1.0, max_timescale: float = 10000.0, dtype: torch.dtype = torch.float32, ): super().__init__() if embedding_dims % 2 != 0: raise ValueError("Embedding dim must be even for RoPE.") self.embedding_dims = embedding_dims self.min_timescale = min_timescale self.max_timescale = max_timescale self.compute_dtype = dtype half_embedding_dim = embedding_dims // 2 fraction = (2.0 * torch.arange(0, half_embedding_dim)) / embedding_dims timescale = (self.min_timescale * (self.max_timescale / self.min_timescale) ** fraction).to(torch.float32) self.register_buffer("timescale", timescale, persistent=False) def forward(self, inputs: torch.Tensor, position: torch.Tensor): """Applies RoPE.""" position = position.unsqueeze(-1).unsqueeze(-1) sinusoid_inp = position / self.timescale sin = torch.sin(sinusoid_inp) cos = torch.cos(sinusoid_inp) first_half, second_half = torch.chunk(inputs.to(torch.float32), 2, dim=-1) first_part = first_half * cos - second_half * sin second_part = second_half * cos + first_half * sin return torch.cat( (first_part.to(self.compute_dtype), second_part.to(self.compute_dtype)), dim=-1, ) def apply_rope(self, inputs: torch.Tensor, sin: torch.Tensor, cos: torch.Tensor): first_half, second_half = torch.chunk(inputs.to(torch.float32), 2, dim=-1) first_part = first_half * cos - second_half * sin second_part = second_half * cos + first_half * sin return torch.cat((first_part.to(self.compute_dtype), second_part.to(self.compute_dtype)), dim=-1) def custom_scaled_dot_product_attention( query: torch.Tensor, key: torch.Tensor, value: torch.Tensor, attn_mask: torch.Tensor | None = None, scale: float = 1.0, is_causal: bool = False, num_gqa_groups: int = 1, ) -> torch.Tensor: """ Custom scaled dot-product attention with GQA support for MPS compatibility. Args: query: (B, N_q, T, H) - Query tensor, N_q = num_query_heads key: (B, N_kv, S, H) - Key tensor, N_kv = num_kv_heads value: (B, N_kv, S, H) - Value tensor attn_mask: (B, 1, T, S) - Attention mask, optional scale: Scaling factor for attention scores is_causal: If True, apply causal masking num_gqa_groups: Number of query groups per KV head (N_q / N_kv) Returns: output: (B, N_q, T, H) - Attention output """ B, N_q, T, H = query.shape _, N_kv, S, _ = key.shape # For GQA, repeat key and value tensors to match query heads if num_gqa_groups > 1: key = key.repeat_interleave(num_gqa_groups, dim=1) # (B, N_q, S, H) value = value.repeat_interleave(num_gqa_groups, dim=1) # (B, N_q, S, H) # Compute attention scores: (B, N_q, T, H) @ (B, N_q, H, S) -> (B, N_q, T, S) scores = torch.matmul(query, key.transpose(-1, -2)) * scale # Apply causal mask if needed if is_causal: causal_mask = torch.tril(torch.ones(T, S, dtype=torch.bool, device=query.device)) scores = scores.masked_fill(~causal_mask, float("-inf")) # Apply attention mask if provided if attn_mask is not None: scores = scores.masked_fill(~attn_mask, float("-inf")) # Softmax over the last dimension (S) attn_weights = F.softmax(scores, dim=-1) # Compute output: (B, N_q, T, S) @ (B, N_q, S, H) -> (B, N_q, T, H) output = torch.matmul(attn_weights, value) return output class CrossAttention(nn.Module): """Cross-Attention using DenseGeneral.""" def __init__( self, config: EncoderConfig | DecoderConfig, q_embed_dim: int, kv_embed_dim: int, num_query_heads: int, num_kv_heads: int, head_dim: int, compute_dtype: torch.dtype, out_embed_dim: int | None = None, ): super().__init__() self.num_query_heads = num_query_heads self.num_kv_heads = num_kv_heads self.head_dim = head_dim self.output_dim = out_embed_dim if out_embed_dim is not None else q_embed_dim self.projected_query_dim = num_query_heads * head_dim if num_query_heads % num_kv_heads != 0: raise ValueError(f"num_query_heads ({num_query_heads}) must be divisible by num_kv_heads ({num_kv_heads})") self.num_gqa_groups = num_query_heads // num_kv_heads # --- Projection Layers using DenseGeneral --- self.q_proj = DenseGeneral( in_shapes=(q_embed_dim,), out_features=(num_query_heads, head_dim), axis=(-1,), weight_dtype=compute_dtype, ) self.k_proj = DenseGeneral( in_shapes=(kv_embed_dim,), out_features=(num_kv_heads, head_dim), axis=(-1,), weight_dtype=compute_dtype, ) self.v_proj = DenseGeneral( in_shapes=(kv_embed_dim,), out_features=(num_kv_heads, head_dim), axis=(-1,), weight_dtype=compute_dtype, ) self.o_proj = DenseGeneral( in_shapes=(num_query_heads, head_dim), out_features=(self.output_dim,), axis=(-2, -1), weight_dtype=compute_dtype, ) # --- Rotary Embedding --- self.rotary_emb = RotaryEmbedding( embedding_dims=self.head_dim, max_timescale=config.rope_theta, dtype=compute_dtype, ) def forward( self, Xq: torch.Tensor, # (B, T, D) T = 1 in AR generation q_positions: torch.Tensor, # (B, T) kv_positions: torch.Tensor | None = None, # (B, S) attn_mask: torch.Tensor | None = None, # None in Decoder Self Attention, Valid mask in Others cache: KVCache | None = None, # None in Encoder, KVCache in Decoder is_causal: bool = False, ) -> tuple[torch.Tensor, tuple[torch.Tensor, torch.Tensor] | None]: """ Performs attention calculation with optional KV caching. Args: Xq: Query tensor (B, T, D). T=1 during single-step decoding. Xkv: Key/Value source tensor (B, S, E). S=1 during single-step decoding for self-attn. q_positions: Positions for queries (B, T). kv_positions: Positions for keys/values (B, S). If None, uses q_positions. attn_mask: Attention mask. cache: KVCache. Returns: A tuple containing: - output: The attention output tensor (B, T, output_dim). - present_kv: The K/V state to be cached for the next step ((B, N, S_new, H), (B, N, S_new, H)). For self-attn, S_new = S_past + S. For cross-attn, S_new = S_kv. """ if kv_positions is None: kv_positions = q_positions original_dtype = Xq.dtype Xq_BxTxNxH = self.q_proj(Xq) Xq_BxNxTxH = Xq_BxTxNxH.transpose(1, 2) attn_k: torch.Tensor | None = cache.k if cache is not None else None attn_v: torch.Tensor | None = cache.v if cache is not None else None # Use custom attention for MPS backend, otherwise use optimized PyTorch function is_mps = Xq.device.type == "mps" and torch.backends.mps.is_available() if is_mps: attn_output = custom_scaled_dot_product_attention( query=Xq_BxNxTxH, key=attn_k, value=attn_v, attn_mask=attn_mask if not is_causal else None, scale=1.0, is_causal=is_causal, num_gqa_groups=self.num_gqa_groups, ) else: attn_output = F.scaled_dot_product_attention( Xq_BxNxTxH, attn_k, attn_v, attn_mask=attn_mask if not is_causal else None, scale=1.0, enable_gqa=self.num_gqa_groups > 1, is_causal=is_causal, ) attn_output = attn_output.transpose(1, 2).contiguous() # (B, T, N, H) output = self.o_proj(attn_output) return output.to(original_dtype) class FusedQKV(nn.Module): def __init__( self, in_features: int, out_features: int, bias: bool = False, num_q_heads: int = 1, q_head_dim: int = 1, num_kv_heads: int = 1, kv_head_dim: int = 1, ): super().__init__() self.num_q_heads = num_q_heads self.q_head_dim = q_head_dim self.num_kv_heads = num_kv_heads self.kv_head_dim = kv_head_dim self.q_output_dim = num_q_heads * q_head_dim self.kv_output_dim = num_kv_heads * kv_head_dim self.linear = nn.Linear(in_features, out_features, bias=bias) def forward(self, inputs: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor]: x = self.linear(inputs) q, k, v = x.split([self.q_output_dim, self.kv_output_dim, self.kv_output_dim], dim=-1) q = q.reshape(q.shape[:-1] + (self.num_q_heads, self.q_head_dim)) k = k.reshape(k.shape[:-1] + (self.num_kv_heads, self.kv_head_dim)) v = v.reshape(v.shape[:-1] + (self.num_kv_heads, self.kv_head_dim)) return q, k, v class SelfAttention(nn.Module): """Attention using DenseGeneral.""" def __init__( self, config: EncoderConfig | DecoderConfig, q_embed_dim: int, kv_embed_dim: int, num_query_heads: int, num_kv_heads: int, head_dim: int, compute_dtype: torch.dtype, out_embed_dim: int | None = None, ): super().__init__() self.num_query_heads = num_query_heads self.num_kv_heads = num_kv_heads self.head_dim = head_dim self.output_dim = out_embed_dim if out_embed_dim is not None else q_embed_dim self.projected_query_dim = num_query_heads * head_dim if num_query_heads % num_kv_heads != 0: raise ValueError(f"num_query_heads ({num_query_heads}) must be divisible by num_kv_heads ({num_kv_heads})") self.num_gqa_groups = num_query_heads // num_kv_heads self.kv_embed_dim = kv_embed_dim self.q_embed_dim = q_embed_dim # --- Projection Layers using DenseGeneral --- self.q_proj = DenseGeneral( in_shapes=(q_embed_dim,), out_features=(num_query_heads, head_dim), axis=(-1,), weight_dtype=compute_dtype, ) self.k_proj = DenseGeneral( in_shapes=(kv_embed_dim,), out_features=(num_kv_heads, head_dim), axis=(-1,), weight_dtype=compute_dtype, ) self.v_proj = DenseGeneral( in_shapes=(kv_embed_dim,), out_features=(num_kv_heads, head_dim), axis=(-1,), weight_dtype=compute_dtype, ) self.o_proj = DenseGeneral( in_shapes=(num_query_heads, head_dim), out_features=(self.output_dim,), axis=(-2, -1), weight_dtype=compute_dtype, ) # --- Rotary Embedding --- self.rotary_emb = RotaryEmbedding( embedding_dims=self.head_dim, max_timescale=config.rope_theta, dtype=compute_dtype, ) self.is_fused_qkv = False def get_linear_weight(self, dense: DenseGeneral): W_dg = dense.weight.data out_features = 1 input_features = 1 for dim in dense.out_features: out_features *= dim for dim in dense.in_shapes: input_features *= dim W_dg_reshaped_for_linear_T = W_dg.reshape(input_features, out_features) linear_weight = W_dg_reshaped_for_linear_T.transpose(0, 1).contiguous() return linear_weight def patch_fused_qkv(self): q_proj_weight = self.get_linear_weight(self.q_proj) k_proj_weight = self.get_linear_weight(self.k_proj) v_proj_weight = self.get_linear_weight(self.v_proj) self.qkv = FusedQKV( self.kv_embed_dim, (self.num_query_heads * self.head_dim + 2 * (self.num_kv_heads * self.head_dim)), bias=False, num_q_heads=self.num_query_heads, q_head_dim=self.head_dim, num_kv_heads=self.num_kv_heads, kv_head_dim=self.head_dim, ) self.qkv.linear.weight.data = torch.cat([q_proj_weight, k_proj_weight, v_proj_weight], dim=0) # print(f"qkv.weight.shape: {self.qkv.linear.weight.shape}") self.is_fused_qkv = True def forward( self, X: torch.Tensor, # (B, T, D) T = 1 in AR generation q_positions: torch.Tensor, # (B, T) kv_positions: torch.Tensor | None = None, # (B, S) attn_mask: torch.Tensor | None = None, # None in Decoder Self Attention, Valid mask in Others cache: KVCache | None = None, # None in Encoder, KVCache in Decoder prefill: bool = False, is_causal: bool = False, current_idx: torch.Tensor | None = None, ) -> tuple[torch.Tensor, tuple[torch.Tensor, torch.Tensor] | None]: """ Performs attention calculation with optional KV caching. Args: Xq: Query tensor (B, T, D). T=1 during single-step decoding. Xkv: Key/Value source tensor (B, S, E). S=1 during single-step decoding for self-attn. q_positions: Positions for queries (B, T). kv_positions: Positions for keys/values (B, S). If None, uses q_positions. attn_mask: Attention mask. cache: KVCache. prefill: If True, use prefill mode. Returns: A tuple containing: - output: The attention output tensor (B, T, output_dim). - present_kv: The K/V state to be cached for the next step ((B, N, S_new, H), (B, N, S_new, H)). For self-attn, S_new = S_past + S. For cross-attn, S_new = S_kv. """ if kv_positions is None: kv_positions = q_positions original_dtype = X.dtype if self.is_fused_qkv: Xq_BxTxNxH, Xk_BxSxKxH, Xv_BxSxKxH = self.qkv(X) else: Xq_BxTxNxH = self.q_proj(X) Xk_BxSxKxH = self.k_proj(X) Xv_BxSxKxH = self.v_proj(X) position = q_positions.unsqueeze(-1).unsqueeze(-1) sinusoid_inp = position / self.rotary_emb.timescale sin = torch.sin(sinusoid_inp) cos = torch.cos(sinusoid_inp) Xq_BxTxNxH = self.rotary_emb.apply_rope(Xq_BxTxNxH, sin, cos) Xk_BxSxKxH = self.rotary_emb.apply_rope(Xk_BxSxKxH, sin, cos) Xq_BxNxTxH = Xq_BxTxNxH.transpose(1, 2) attn_k: torch.Tensor | None = cache.k if cache is not None else None attn_v: torch.Tensor | None = cache.v if cache is not None else None Xk_BxKxSxH = Xk_BxSxKxH.transpose(1, 2) # (B, K, S, H) Xv_BxKxSxH = Xv_BxSxKxH.transpose(1, 2) # (B, K, S, H) if cache is None: attn_k = Xk_BxKxSxH attn_v = Xv_BxKxSxH elif prefill: attn_k, attn_v = Xk_BxKxSxH, Xv_BxKxSxH cache.prefill(attn_k, attn_v) else: attn_k, attn_v = cache.update(Xk_BxKxSxH, Xv_BxKxSxH, current_idx) # Use custom attention for MPS backend, otherwise use optimized PyTorch function is_mps = Xv_BxSxKxH.device.type == "mps" and torch.backends.mps.is_available() if is_mps: attn_output = custom_scaled_dot_product_attention( query=Xq_BxNxTxH, key=attn_k, value=attn_v, attn_mask=attn_mask if not is_causal else None, scale=1.0, is_causal=is_causal, num_gqa_groups=self.num_gqa_groups, ) else: attn_output = F.scaled_dot_product_attention( Xq_BxNxTxH, attn_k, attn_v, attn_mask=attn_mask if not is_causal else None, scale=1.0, enable_gqa=self.num_gqa_groups > 1, is_causal=is_causal, ) attn_output = attn_output.transpose(1, 2).contiguous() # (B, T, N, H) output = self.o_proj(attn_output) return output.to(original_dtype) class EncoderLayer(nn.Module): """Transformer Encoder Layer using DenseGeneral.""" def __init__(self, config: DiaConfig, compute_dtype: torch.dtype): super().__init__() self.config = config enc_config = config.encoder_config embed_dim = enc_config.hidden_size self.compute_dtype = compute_dtype self.pre_sa_norm = RMSNorm( embed_dim, eps=enc_config.norm_eps, dtype=torch.float32, ) self.self_attention = SelfAttention( enc_config, q_embed_dim=embed_dim, kv_embed_dim=embed_dim, num_query_heads=enc_config.num_attention_heads, num_kv_heads=enc_config.num_key_value_heads, head_dim=enc_config.head_dim, compute_dtype=compute_dtype, out_embed_dim=embed_dim, ) self.post_sa_norm = RMSNorm( embed_dim, eps=enc_config.norm_eps, dtype=torch.float32, ) self.mlp = MlpBlock( embed_dim=embed_dim, intermediate_dim=enc_config.intermediate_size, compute_dtype=compute_dtype, ) def forward( self, x: torch.Tensor, state: EncoderInferenceState, ) -> torch.Tensor: residual = x x_norm = self.pre_sa_norm(x).to(self.compute_dtype) sa_out = self.self_attention( X=x_norm, q_positions=state.positions, kv_positions=state.positions, attn_mask=state.attn_mask, ) x = residual + sa_out residual = x x_norm = self.post_sa_norm(x).to(self.compute_dtype) mlp_out = self.mlp(x_norm) x = residual + mlp_out return x class Encoder(nn.Module): """Transformer Encoder Stack using DenseGeneral.""" def __init__(self, config: DiaConfig, compute_dtype: torch.dtype): super().__init__() self.config = config enc_config = config.encoder_config self.compute_dtype = compute_dtype self.embedding = nn.Embedding( enc_config.vocab_size, enc_config.hidden_size, dtype=compute_dtype, ) self.layers = nn.ModuleList([EncoderLayer(config, compute_dtype) for _ in range(enc_config.num_hidden_layers)]) self.norm = RMSNorm( enc_config.hidden_size, eps=enc_config.norm_eps, dtype=torch.float32, ) def forward( self, x_ids: torch.Tensor, state: EncoderInferenceState, ) -> torch.Tensor: x = self.embedding(x_ids) for layer in self.layers: x = layer(x, state) x = self.norm(x).to(self.compute_dtype) return x class DecoderLayer(nn.Module): """Transformer Decoder Layer using DenseGeneral.""" def __init__(self, config: DiaConfig, compute_dtype: torch.dtype): super().__init__() self.config = config dec_config = config.decoder_config enc_config = config.encoder_config dec_embed_dim = dec_config.hidden_size enc_embed_dim = enc_config.hidden_size self.compute_dtype = compute_dtype # Norms self.pre_sa_norm = RMSNorm( dec_embed_dim, eps=dec_config.norm_eps, dtype=torch.float32, ) self.pre_ca_norm = RMSNorm( dec_embed_dim, eps=dec_config.norm_eps, dtype=torch.float32, ) self.pre_mlp_norm = RMSNorm( dec_embed_dim, eps=dec_config.norm_eps, dtype=torch.float32, ) # Self-Attention (GQA) with Causal Masking self.self_attention = SelfAttention( dec_config, q_embed_dim=dec_embed_dim, kv_embed_dim=dec_embed_dim, num_query_heads=dec_config.num_attention_heads, num_kv_heads=dec_config.num_key_value_heads, head_dim=dec_config.head_dim, compute_dtype=compute_dtype, out_embed_dim=dec_embed_dim, ) # Cross-Attention (MHA) self.cross_attention = CrossAttention( dec_config, q_embed_dim=dec_embed_dim, kv_embed_dim=enc_embed_dim, # Note kv_embed_dim num_query_heads=dec_config.cross_num_attention_heads, num_kv_heads=dec_config.cross_num_key_value_heads, head_dim=dec_config.cross_head_dim, compute_dtype=compute_dtype, out_embed_dim=dec_embed_dim, ) # MLP self.mlp = MlpBlock( embed_dim=dec_embed_dim, intermediate_dim=dec_config.intermediate_size, compute_dtype=compute_dtype, ) def forward( self, x: torch.Tensor, state: DecoderInferenceState, self_attn_cache: KVCache | None = None, cross_attn_cache: KVCache | None = None, prefill: bool = False, current_idx: int = 0, ) -> torch.Tensor: residual = x x_norm = self.pre_sa_norm(x).to(self.compute_dtype) self_attn_mask = state.casual_attn_mask[None, None, current_idx] sa_out = self.self_attention( X=x_norm, # (2, 1, D) q_positions=state.dec_positions, # (2, 1) kv_positions=state.dec_positions, # (2, 1) attn_mask=self_attn_mask, cache=self_attn_cache, prefill=prefill, is_causal=prefill, current_idx=current_idx, ) x = residual + sa_out residual = x x_norm = self.pre_ca_norm(x).to(self.compute_dtype) ca_out = self.cross_attention( Xq=x_norm, q_positions=state.dec_positions, kv_positions=state.enc_positions, attn_mask=state.cross_attn_mask, cache=cross_attn_cache, ) x = residual + ca_out residual = x x_norm = self.pre_mlp_norm(x).to(self.compute_dtype) mlp_out = self.mlp(x_norm) x = residual + mlp_out return x class Decoder(nn.Module): """Transformer Decoder Stack using DenseGeneral.""" def __init__(self, config: DiaConfig, compute_dtype: torch.dtype): super().__init__() self.config = config dec_config = config.decoder_config self.num_channels = dec_config.num_channels self.num_layers = dec_config.num_hidden_layers self.embeddings = nn.ModuleList( [ nn.Embedding(dec_config.vocab_size, dec_config.hidden_size, dtype=compute_dtype) for _ in range(self.num_channels) ] ) self.layers = nn.ModuleList( [DecoderLayer(config=config, compute_dtype=compute_dtype) for _ in range(self.num_layers)] ) self.norm = RMSNorm( dec_config.hidden_size, eps=dec_config.norm_eps, dtype=torch.float32, ) self.logits_dense = DenseGeneral( in_shapes=(dec_config.hidden_size,), out_features=(self.num_channels, dec_config.vocab_size), axis=(-1,), weight_dtype=compute_dtype, ) def precompute_cross_attn_cache( self, enc_out: torch.Tensor, # (B, S, E) ) -> list[KVCache]: """ Computes the Key and Value tensors for cross-attention for each layer from the encoder output. """ per_layer_kv_cache: list[KVCache] = [] for layer in self.layers: cross_attn_module = layer.cross_attention k_proj = cross_attn_module.k_proj(enc_out) v_proj = cross_attn_module.v_proj(enc_out) k = k_proj.transpose(1, 2) v = v_proj.transpose(1, 2) per_layer_kv_cache.append(KVCache.from_kv(k, v)) return per_layer_kv_cache def decode_step( self, tgt_ids_Bx1xC: torch.Tensor, # [B, 1, C] state: DecoderInferenceState, current_idx: int, ) -> torch.Tensor: """ Performs a single decoding step, managing KV caches layer by layer. Returns: A tuple containing: - logits_Bx1xCV: The final output logits for the current step (B, 1, C*V), cast to float32. """ x = None for i in range(self.num_channels): channel_tokens = tgt_ids_Bx1xC[..., i] channel_embed = self.embeddings[i](channel_tokens) x = channel_embed if x is None else x + channel_embed for i, layer in enumerate(self.layers): self_cache = state.self_attn_cache[i] cross_cache = state.cross_attn_cache[i] x = layer( x, # (2, 1, D) state, self_attn_cache=self_cache, cross_attn_cache=cross_cache, current_idx=current_idx, ) x = self.norm(x) logits_Bx1xCxV = self.logits_dense(x) return logits_Bx1xCxV.to(torch.float32) def forward(self, tgt_ids_BxTxC: torch.Tensor, state: DecoderInferenceState) -> torch.Tensor: """ Forward pass for the Decoder stack, managing KV caches. Args: tgt_ids_BxTxC: Target token IDs (B, T, C). encoder_out: Output from the encoder (B, S, E). tgt_positions: Positions for target sequence (B, T). src_positions: Positions for source sequence (B, S). self_attn_mask: Mask for self-attention. cross_attn_mask: Mask for cross-attention. past_key_values: List containing the self-attention KV cache for each layer from the previous decoding step. `len(past_key_values)` should equal `num_layers`. precomputed_cross_attn_kv: A single tuple containing the pre-computed K/V cache derived from `encoder_out`. This is passed identically to all layers. Returns: A tuple containing: - logits: The final output logits (B, T, C * V), cast to float32. - present_key_values: A list containing the updated self-attention KV cache for each layer for the *current* decoding step. """ _, _, num_channels_in = tgt_ids_BxTxC.shape assert num_channels_in == self.num_channels, "Input channels mismatch" # Embeddings x = None for i in range(self.num_channels): channel_tokens = tgt_ids_BxTxC[..., i] channel_embed = self.embeddings[i](channel_tokens) x = channel_embed if x is None else x + channel_embed for i, layer in enumerate(self.layers): self_cache = state.self_attn_cache[i] cross_cache = state.cross_attn_cache[i] x = layer( x, state, self_attn_cache=self_cache, cross_attn_cache=cross_cache, prefill=True, ) # Final Norm x = self.norm(x) logits_BxTxCxV = self.logits_dense(x) return logits_BxTxCxV.to(torch.float32) class DiaModel( nn.Module, PyTorchModelHubMixin, repo_url="https://github.com/nari-labs/dia", pipeline_tag="text-to-speech", license="apache-2.0", coders={ DiaConfig: ( lambda x: x.model_dump(), lambda data: DiaConfig.model_validate(data), ), }, ): """PyTorch Dia Model using DenseGeneral.""" def __init__(self, config: DiaConfig, compute_dtype: torch.dtype): super().__init__() self.config = config self.encoder = Encoder(config, compute_dtype) self.decoder = Decoder(config, compute_dtype) ================================================ FILE: dia/model.py ================================================ import time from enum import Enum from typing import Callable import numpy as np import torch import torch.nn.functional as F import torchaudio from .audio import apply_audio_delay, build_delay_indices, build_revert_indices, revert_audio_delay from .config import DiaConfig from .layers import DiaModel from .state import DecoderInferenceState, DecoderOutput, EncoderInferenceState DEFAULT_SAMPLE_RATE = 44100 SAMPLE_RATE_RATIO = 512 def _get_default_device(): if torch.cuda.is_available(): return torch.device("cuda") elif hasattr(torch.backends, "mps") and torch.backends.mps.is_available(): return torch.device("mps") return torch.device("cpu") def _sample_next_token( logits_BCxV: torch.Tensor, temperature: float, top_p: float, top_k: int | None, audio_eos_value: int, ) -> torch.Tensor: if temperature == 0.0: return torch.argmax(logits_BCxV, dim=-1) logits_BCxV = logits_BCxV / temperature if audio_eos_value is not None and audio_eos_value >= 0: top_logit_indices_BC = torch.argmax(logits_BCxV, dim=-1) eos_not_highest_mask_BC = top_logit_indices_BC != audio_eos_value mask_eos_unless_highest_BCxV = torch.zeros_like(logits_BCxV, dtype=torch.bool) mask_eos_unless_highest_BCxV[eos_not_highest_mask_BC, audio_eos_value] = True logits_BCxV = logits_BCxV.masked_fill(mask_eos_unless_highest_BCxV, -torch.inf) eos_highest_mask_BC = top_logit_indices_BC == audio_eos_value mask_eos_highest_BCxV = torch.zeros_like(logits_BCxV, dtype=torch.bool) mask_eos_highest_BCxV[eos_highest_mask_BC, :audio_eos_value] = True logits_BCxV = logits_BCxV.masked_fill(mask_eos_highest_BCxV, -torch.inf) if top_k is not None: _, top_k_indices_BCxV = torch.topk(logits_BCxV, k=top_k, dim=-1) mask = torch.ones_like(logits_BCxV, dtype=torch.bool) mask = mask.scatter(dim=-1, index=top_k_indices_BCxV, value=False) logits_BCxV = logits_BCxV.masked_fill(mask, -torch.inf) if top_p < 1.0: probs_BCxV = torch.softmax(logits_BCxV, dim=-1) sorted_probs_BCxV, sorted_indices_BCxV = torch.sort(probs_BCxV, dim=-1, descending=True) cumulative_probs_BCxV = torch.cumsum(sorted_probs_BCxV, dim=-1) sorted_indices_to_remove_BCxV = cumulative_probs_BCxV > top_p sorted_indices_to_remove_BCxV = torch.roll(sorted_indices_to_remove_BCxV, shifts=1, dims=-1) sorted_indices_to_remove_BCxV[..., 0] = torch.zeros_like(sorted_indices_to_remove_BCxV[..., 0]) indices_to_remove_BCxV = torch.zeros_like(sorted_indices_to_remove_BCxV) indices_to_remove_BCxV = indices_to_remove_BCxV.scatter( dim=-1, index=sorted_indices_BCxV, src=sorted_indices_to_remove_BCxV ) logits_BCxV = logits_BCxV.masked_fill(indices_to_remove_BCxV, -torch.inf) final_probs_BCxV = torch.softmax(logits_BCxV, dim=-1) sampled_indices_BC = torch.multinomial(final_probs_BCxV, num_samples=1) sampled_indices_C = sampled_indices_BC.squeeze(-1) return sampled_indices_C class ComputeDtype(str, Enum): FLOAT32 = "float32" FLOAT16 = "float16" BFLOAT16 = "bfloat16" def to_dtype(self) -> torch.dtype: if self == ComputeDtype.FLOAT32: return torch.float32 elif self == ComputeDtype.FLOAT16: return torch.float16 elif self == ComputeDtype.BFLOAT16: return torch.bfloat16 else: raise ValueError(f"Unsupported compute dtype: {self}") class Dia: def __init__( self, config: DiaConfig, compute_dtype: str | ComputeDtype = ComputeDtype.FLOAT32, device: torch.device | None = None, load_dac: bool = True, ): """Initializes the Dia model. Args: config: The configuration object for the model. compute_dtype: The computation dtype to use. device: The device to load the model onto. If None, will automatically select the best available device. load_dac: Whether to load the DAC model. Raises: RuntimeError: If there is an error loading the DAC model. """ super().__init__() self.config = config self.device = device if device is not None else _get_default_device() if isinstance(compute_dtype, str): compute_dtype = ComputeDtype(compute_dtype) self.compute_dtype = compute_dtype.to_dtype() self.model: DiaModel = DiaModel(config, self.compute_dtype) self.dac_model = None self._compiled_step = None self.load_dac = load_dac if not self.load_dac: print("Warning: DAC model will not be loaded. This is not recommended.") if torch.cuda.is_available(): torch.backends.cuda.matmul.allow_tf32 = True @classmethod def from_local( cls, config_path: str, checkpoint_path: str, compute_dtype: str | ComputeDtype = ComputeDtype.FLOAT32, device: torch.device | None = None, load_dac: bool = True, ) -> "Dia": """Loads the Dia model from local configuration and checkpoint files. Args: config_path: Path to the configuration JSON file. checkpoint_path: Path to the model checkpoint (.pth) file. compute_dtype: The computation dtype to use. device: The device to load the model onto. If None, will automatically select the best available device. load_dac: Whether to load the DAC model. Returns: An instance of the Dia model loaded with weights and set to eval mode. Raises: FileNotFoundError: If the config or checkpoint file is not found. RuntimeError: If there is an error loading the checkpoint. """ config = DiaConfig.load(config_path) if config is None: raise FileNotFoundError(f"Config file not found at {config_path}") dia = cls(config, compute_dtype, device, load_dac) try: state_dict = torch.load(checkpoint_path, map_location=dia.device) dia.model.load_state_dict(state_dict) except FileNotFoundError: raise FileNotFoundError(f"Checkpoint file not found at {checkpoint_path}") except Exception as e: raise RuntimeError(f"Error loading checkpoint from {checkpoint_path}") from e dia.model.to(dia.device) dia.model.eval() if load_dac: dia._load_dac_model() return dia @classmethod def from_pretrained( cls, model_name: str = "nari-labs/Dia-1.6B-0626", compute_dtype: str | ComputeDtype = ComputeDtype.FLOAT32, device: torch.device | None = None, load_dac: bool = True, ) -> "Dia": """Loads the Dia model from a Hugging Face Hub repository. Downloads the configuration and checkpoint files from the specified repository ID and then loads the model. Args: model_name: The Hugging Face Hub repository ID (e.g., "nari-labs/Dia-1.6B-0626"). compute_dtype: The computation dtype to use. device: The device to load the model onto. If None, will automatically select the best available device. load_dac: Whether to load the DAC model. Returns: An instance of the Dia model loaded with weights and set to eval mode. Raises: FileNotFoundError: If config or checkpoint download/loading fails. RuntimeError: If there is an error loading the checkpoint. """ if isinstance(compute_dtype, str): compute_dtype = ComputeDtype(compute_dtype) # Load model directly using DiaModel's from_pretrained which handles HF download try: loaded_model = DiaModel.from_pretrained(model_name, compute_dtype=compute_dtype.to_dtype()) except Exception as e: raise RuntimeError(f"Error loading model from Hugging Face Hub ({model_name})") from e config = loaded_model.config # Get config from the loaded model dia = cls(config, compute_dtype, device, load_dac) dia.model = loaded_model # Assign the already loaded model dia.model.to(dia.device) dia.model.eval() if load_dac: dia._load_dac_model() return dia def _load_dac_model(self): """Loads the Descript Audio Codec (DAC) model. Downloads the DAC model if necessary and loads it onto the specified device. Sets the DAC model to evaluation mode. Raises: RuntimeError: If downloading or loading the DAC model fails. """ import dac try: dac_model_path = dac.utils.download() dac_model = dac.DAC.load(dac_model_path).to(self.device) dac_model.eval() # Ensure DAC is in eval mode except Exception as e: raise RuntimeError("Failed to load DAC model") from e self.dac_model = dac_model def _encode_text(self, text: str) -> torch.Tensor: """Encodes the input text string into a tensor of token IDs using byte-level encoding. Special tokens [S1] and [S2] are replaced by their byte values. The resulting sequence is truncated to the maximum configured text length. Args: text: The input text string. Returns: A tensor containing the encoded byte token IDs. """ max_len = self.config.encoder_config.max_position_embeddings byte_text = text.encode("utf-8") # Replace special tokens with their byte values if needed by the specific tokenizer/config # Assuming byte values 1 and 2 are correct placeholders based on original code replaced_bytes = byte_text.replace(b"[S1]", b"\x01").replace(b"[S2]", b"\x02") text_tokens = list(replaced_bytes) return torch.tensor( text_tokens[:max_len], dtype=torch.long, device=self.device, ) def _pad_text_input(self, text_tokens: list[torch.Tensor]) -> torch.Tensor: """Pads the text input to the maximum length.""" text_pad_value = 0 max_len = self.config.encoder_config.max_position_embeddings batch_size = len(text_tokens) src_tokens = torch.full( (batch_size, 1, max_len), fill_value=text_pad_value, dtype=torch.long, device=self.device, ) for i in range(batch_size): current_len = len(text_tokens[i]) src_tokens[i, 0, :current_len] = text_tokens[i] return src_tokens def _prepare_audio_prompt(self, audio_prompts: list[torch.Tensor | None]) -> tuple[torch.Tensor, list[int]]: """Prepares the audio prompt tensor for the decoder. Handles padding, adds the beginning-of-sequence (BOS) token, applies the delay pattern, and determines the number of prefill steps for each item in the batch. Args: audio_prompts: A list of audio prompt tensors (encoded DAC frames) or None. Each tensor should have shape [T, C]. Returns: A tuple containing: - delayed_batch (torch.Tensor): The prepared audio prompt tensor with delays applied, shape [B, T_max_padded, C]. - prefill_steps (list[int]): A list containing the number of valid tokens (including BOS) for each prompt in the batch. """ num_channels = self.config.decoder_config.num_channels audio_bos_value = self.config.bos_token_id delay_pattern = self.config.delay_pattern max_delay_pattern = max(delay_pattern) batch_size = len(audio_prompts) max_len = max(p.shape[0] if p is not None else 0 for p in audio_prompts) + max_delay_pattern prefill_steps = [] prefill = torch.full( (batch_size, max_len, num_channels), fill_value=-1, dtype=torch.int, device=self.device, ) prefill[:, 0, :] = audio_bos_value for i in range(batch_size): prompt = audio_prompts[i] if prompt is not None: prompt = prompt.to(device=self.device, dtype=torch.int) prefill[i, 1 : prompt.shape[0] + 1, :] = prompt prefill_steps.append(prompt.shape[0] + 1) else: prefill_steps.append(1) delay_precomp = build_delay_indices( B=batch_size, T=max_len, C=num_channels, delay_pattern=delay_pattern, ) delayed_batch = apply_audio_delay( audio_BxTxC=prefill, pad_value=-1, bos_value=audio_bos_value, precomp=delay_precomp, ) return delayed_batch, prefill_steps def _prepare_generation( self, text: torch.Tensor, audio_prompts: list[torch.Tensor | None], max_tokens: int | None = None, attn_fn: Callable = F.scaled_dot_product_attention, ): """Initializes the model state for generation. Encodes the text input (conditional and unconditional), prepares the encoder and decoder states (including KV caches and cross-attention), prepares the audio prompt, and performs the initial decoder prefill steps based on the audio prompts. Args: text: The padded text input tensor, shape [B, 1, T_text]. audio_prompts: A list of prepared audio prompt tensors or None. Returns: A tuple containing: - dec_state (DecoderInferenceState): The initialized decoder state. - dec_output (DecoderOutput): The initialized decoder output manager, containing the prefilled audio tokens. """ batch_size = text.shape[0] enc_input_uncond = torch.zeros_like(text) enc_input_cond = text stacked_inputs = torch.stack([enc_input_uncond, enc_input_cond], dim=1) enc_input = stacked_inputs.view(2 * batch_size, -1) enc_state = EncoderInferenceState.new(self.config, enc_input_cond) encoder_out = self.model.encoder(enc_input, enc_state) dec_cross_attn_cache = self.model.decoder.precompute_cross_attn_cache(encoder_out) dec_state = DecoderInferenceState.new( self.config, enc_state, encoder_out, dec_cross_attn_cache, self.compute_dtype, max_generation_length=max_tokens, ) prefill, prefill_steps = self._prepare_audio_prompt(audio_prompts) dec_output = DecoderOutput.new(batch_size, self.config, self.device) dec_output.prefill(prefill, prefill_steps) dec_step = min(prefill_steps) - 1 if dec_step > 0: dec_state.prepare_step(0, dec_step) tokens_BxTxC = dec_output.get_tokens_at(0, dec_step).repeat_interleave(2, dim=0) self.model.decoder.forward(tokens_BxTxC, dec_state) return dec_state, dec_output def _decoder_step( self, tokens_Bx1xC: torch.Tensor, dec_state: DecoderInferenceState, cfg_scale: float, temperature: float, top_p: float, top_k: int, current_idx: int, ) -> torch.Tensor: """Performs a single step of the decoder inference. Takes the tokens from the previous step, runs them through the decoder (for both conditional and unconditional paths), applies classifier-free guidance (CFG), samples the next token using temperature, top-p, and top-k sampling, and applies constraints (e.g., preventing EOS in certain channels). Args: tokens_Bx1xC: The input tokens for the current step, shape [2*B, 1, C]. Repeated for CFG (unconditional and conditional). dec_state: The current state of the decoder (KV caches, etc.). cfg_scale: The scale factor for classifier-free guidance. temperature: The temperature for sampling. top_p: The cumulative probability threshold for top-p sampling. top_k: The number of top logits to consider for top-k sampling. current_idx: The current generation step index. Returns: torch.Tensor: The sampled next tokens for each item in the batch, shape [B, C]. """ B = tokens_Bx1xC.shape[0] // 2 audio_eos_value = self.config.eos_token_id logits_Bx1xCxV = self.model.decoder.decode_step(tokens_Bx1xC, dec_state, current_idx) logits_last_2BxCxV = logits_Bx1xCxV[:, -1] logits_last_Bx2xCxV = logits_last_2BxCxV.view(B, 2, *logits_last_2BxCxV.shape[1:]) uncond_logits_BxCxV = logits_last_Bx2xCxV[:, 0, :, :] # Shape [B, C, V] cond_logits_BxCxV = logits_last_Bx2xCxV[:, 1, :, :] # Shape [B, C, V] logits_BxCxV = cond_logits_BxCxV + cfg_scale * (cond_logits_BxCxV - uncond_logits_BxCxV) _, top_k_indices_BxCxk = torch.topk(logits_BxCxV, k=top_k, dim=-1) mask_BxCxV = torch.ones_like(logits_BxCxV, dtype=torch.bool) mask_BxCxV = mask_BxCxV.scatter(dim=-1, index=top_k_indices_BxCxk, value=False) logits_BxCxV = cond_logits_BxCxV.masked_fill(mask_BxCxV, -torch.inf) logits_BxCxV[:, :, audio_eos_value + 1 :] = torch.full_like( logits_BxCxV[:, :, audio_eos_value + 1 :], fill_value=-torch.inf, ) logits_BxCxV[:, 1:, audio_eos_value:] = torch.full_like( logits_BxCxV[:, 1:, audio_eos_value:], fill_value=-torch.inf, ) flat_logits_BCxV = logits_BxCxV.view(B * self.config.decoder_config.num_channels, -1) pred_BC = _sample_next_token( flat_logits_BCxV.float(), temperature=temperature, top_p=top_p, top_k=top_k, audio_eos_value=audio_eos_value, ) pred_BxC = pred_BC.view(B, self.config.decoder_config.num_channels) return pred_BxC def _generate_output(self, generated_codes: torch.Tensor, lengths_Bx: torch.Tensor) -> list[np.ndarray]: """Converts generated delayed codes into audio waveforms. Reverts the delay pattern applied during generation, decodes the resulting codebook using the DAC model (if loaded), and returns a list of audio waveforms as NumPy arrays. If DAC is not loaded, returns the raw codebook indices. Args: generated_codes: The tensor of generated audio codes with delays, shape [B, T_gen, C]. lengths_Bx: A tensor containing the valid length of generated codes (excluding padding and BOS/EOS markers) for each item in the batch, shape [B]. Returns: A list of NumPy arrays, where each array represents the generated audio waveform for one item in the batch. If DAC is not loaded, returns the raw, reverted codebook indices as NumPy arrays. """ num_channels = self.config.decoder_config.num_channels batch_size = generated_codes.shape[0] seq_length = generated_codes.shape[1] delay_pattern = self.config.delay_pattern audio_pad_value = self.config.pad_token_id max_delay_pattern = max(delay_pattern) revert_precomp = build_revert_indices( B=batch_size, T=seq_length, C=num_channels, delay_pattern=delay_pattern, ) codebook = revert_audio_delay( audio_BxTxC=generated_codes, pad_value=audio_pad_value, precomp=revert_precomp, T=seq_length, )[:, :-max_delay_pattern, :] min_valid_index = 0 max_valid_index = 1023 invalid_mask = (codebook < min_valid_index) | (codebook > max_valid_index) codebook[invalid_mask] = 0 audios = [] if self.load_dac: for i in range(batch_size): audio = self._decode(codebook[i, : lengths_Bx[i], :]) audio_np = audio.cpu().numpy() audios.append(audio_np) else: for i in range(batch_size): audios.append(codebook[i, : lengths_Bx[i], :].cpu().numpy()) return audios @torch.no_grad() @torch.inference_mode() def _encode(self, audio: torch.Tensor) -> torch.Tensor: """ Encodes the given audio waveform into a tensor of DAC codebook indices """ audio = audio.unsqueeze(0) audio_data = self.dac_model.preprocess(audio, DEFAULT_SAMPLE_RATE) _, encoded_frame, _, _, _ = self.dac_model.encode(audio_data) encoded_frame: torch.Tensor return encoded_frame.squeeze(0).transpose(0, 1) @torch.no_grad() @torch.inference_mode() def _decode(self, audio_codes: torch.Tensor) -> torch.Tensor: """ Decodes the given frames into an output audio waveform """ audio_codes = audio_codes.unsqueeze(0).transpose(1, 2) audio_values, _, _ = self.dac_model.quantizer.from_codes(audio_codes) audio_values = self.dac_model.decode(audio_values) audio_values: torch.Tensor return audio_values.squeeze() def load_audio(self, audio_path: str) -> torch.Tensor: """Loads and preprocesses an audio file for use as a prompt. Loads the audio file, resamples it to the target sample rate if necessary, preprocesses it using the DAC model's preprocessing, and encodes it into DAC codebook indices. Args: audio_path: Path to the audio file. Returns: torch.Tensor: The encoded audio prompt as DAC codebook indices, shape [T, C]. Raises: RuntimeError: If the DAC model is not loaded (`load_dac=False` during init). FileNotFoundError: If the audio file cannot be found. Exception: If there's an error during loading or processing. """ if self.dac_model is None: raise RuntimeError("DAC model is required for loading audio prompts but was not loaded.") audio, sr = torchaudio.load(audio_path, channels_first=True) # C, T if sr != DEFAULT_SAMPLE_RATE: audio = torchaudio.functional.resample(audio, sr, DEFAULT_SAMPLE_RATE) # Convert to mono if stereo if audio.shape[0] > 1: audio = torch.mean(audio, dim=0, keepdim=True) # Average channels to get mono return self._encode(audio.to(self.device)) def save_audio(self, path: str, audio: np.ndarray): """Saves the generated audio waveform to a file. Uses the soundfile library to write the NumPy audio array to the specified path with the default sample rate. Args: path: The path where the audio file will be saved. audio: The audio waveform as a NumPy array. """ import soundfile as sf sf.write(path, audio, DEFAULT_SAMPLE_RATE) @torch.inference_mode() def generate( self, text: str | list[str], max_tokens: int = 3072, cfg_scale: float = 3.0, temperature: float = 1.2, top_p: float = 0.95, use_torch_compile: bool = False, cfg_filter_top_k: int = 45, audio_prompt: list[str | torch.Tensor | None] | str | torch.Tensor | None = None, audio_prompt_path: list[str | torch.Tensor | None] | str | torch.Tensor | None = None, use_cfg_filter: bool | None = None, verbose: bool = False, ) -> np.ndarray | list[np.ndarray]: """Generates audio corresponding to the input text. Args: text: The input text prompt, or a list of text prompts for batch generation. max_tokens: The maximum number of audio tokens to generate per prompt. Defaults to the model's configured audio length if None. cfg_scale: The scale factor for classifier-free guidance (CFG). Higher values lead to stronger guidance towards the text prompt. temperature: The temperature for sampling. Higher values increase randomness. top_p: The cumulative probability threshold for nucleus (top-p) sampling. use_torch_compile: Whether to compile the generation steps using torch.compile. Can significantly speed up generation after the initial compilation overhead. Defaults to False. cfg_filter_top_k: The number of top logits to consider during CFG filtering. (Note: This parameter name might be slightly misleading based on the code; it's used in the `_sample_next_token` function.) audio_prompt: An audio prompt or list of prompts to condition the generation. Can be a file path (str), a pre-loaded tensor (DAC codes), or None. If a list, its length must match the batch size of the text input. audio_prompt_path: (Deprecated) Use `audio_prompt` instead. use_cfg_filter: (Deprecated) This parameter is no longer used. verbose: If True, prints progress information during generation, including speed metrics. Returns: If a single text prompt was provided, returns a NumPy array containing the generated audio waveform. If a list of text prompts was provided, returns a list of NumPy arrays, each corresponding to a prompt in the input list. Returns None for a sequence if no audio was generated for it. """ batch_size = len(text) if isinstance(text, list) else 1 audio_eos_value = self.config.eos_token_id audio_pad_value = self.config.pad_token_id delay_pattern = self.config.delay_pattern max_delay_pattern = max(delay_pattern) delay_pattern_Cx = torch.tensor(delay_pattern, device=self.device, dtype=torch.long) self.model.eval() if audio_prompt_path: print("Warning: audio_prompt_path is deprecated. Use audio_prompt instead.") audio_prompt = audio_prompt_path if use_cfg_filter is not None: print("Warning: use_cfg_filter is deprecated.") if verbose: total_start_time = time.time() if use_torch_compile and not hasattr(self, "_compiled"): # Compilation can take about a minute. self._prepare_generation = torch.compile(self._prepare_generation, dynamic=True, fullgraph=True) self._decoder_step = torch.compile(self._decoder_step, fullgraph=True, mode="max-autotune") self._compiled = True if isinstance(audio_prompt, list): audio_prompt = [self.load_audio(p) if isinstance(p, str) else p for p in audio_prompt] elif isinstance(audio_prompt, str): audio_prompt = [self.load_audio(audio_prompt)] elif isinstance(audio_prompt, torch.Tensor): audio_prompt = [audio_prompt] elif audio_prompt is None: audio_prompt = [None] * batch_size assert len(audio_prompt) == batch_size, "Number of audio prompts must match batch size" if isinstance(text, list): text = [self._encode_text(t) for t in text] else: text = [self._encode_text(text)] text = self._pad_text_input(text) dec_state, dec_output = self._prepare_generation(text, audio_prompt, max_tokens=max_tokens) dec_step = min(dec_output.prefill_steps) - 1 current_idx = torch.tensor([dec_step], device=self.device) eos_detected_Bx = torch.zeros((batch_size,), dtype=torch.bool, device=self.device) eos_countdown_Bx = torch.full((batch_size,), -1, dtype=torch.long, device=self.device) finished_step_Bx = torch.full((batch_size,), -1, dtype=torch.long, device=self.device) bos_over = False if verbose: print("generate: starting generation loop") if use_torch_compile: print("generate: using use_torch_compile=True, the first step may be slow") start_time = time.time() # --- Generation Loop --- while dec_step < max_tokens: if (eos_countdown_Bx == 0).all(): break current_step_idx = dec_step + 1 torch.compiler.cudagraph_mark_step_begin() dec_state.prepare_step(dec_step) tokens_Bx1xC = dec_output.get_tokens_at(dec_step).repeat_interleave(2, dim=0) # Repeat for CFG pred_BxC = self._decoder_step( tokens_Bx1xC, dec_state, cfg_scale, temperature, top_p, cfg_filter_top_k, current_idx, ) current_idx += 1 active_mask_Bx = eos_countdown_Bx != 0 eos_trigger_Bx = torch.zeros_like(active_mask_Bx) if active_mask_Bx.any(): is_eos_token = (~eos_detected_Bx[active_mask_Bx]) & (pred_BxC[active_mask_Bx, 0] == audio_eos_value) is_max_len = current_step_idx >= max_tokens - max_delay_pattern eos_trigger_Bx[active_mask_Bx] = is_eos_token | is_max_len eos_detected_Bx |= eos_trigger_Bx start_countdown_mask_Bx = eos_trigger_Bx & (eos_countdown_Bx < 0) if start_countdown_mask_Bx.any(): eos_countdown_Bx[start_countdown_mask_Bx] = max_delay_pattern finished_step_Bx[start_countdown_mask_Bx] = current_step_idx padding_mask_Bx = eos_countdown_Bx > 0 if padding_mask_Bx.any(): pred_active_BxC = pred_BxC[padding_mask_Bx].clone() countdown_active_Bx = eos_countdown_Bx[padding_mask_Bx] step_after_eos_Bx = max_delay_pattern - countdown_active_Bx step_after_eos_Bx_ = step_after_eos_Bx.unsqueeze(1) delay_pattern_Cx_ = delay_pattern_Cx.unsqueeze(0) eos_mask_NxC = step_after_eos_Bx_ == delay_pattern_Cx_ pad_mask_NxC = step_after_eos_Bx_ > delay_pattern_Cx_ pred_active_BxC[eos_mask_NxC] = audio_eos_value pred_active_BxC[pad_mask_NxC] = audio_pad_value pred_BxC[padding_mask_Bx] = pred_active_BxC eos_countdown_Bx[padding_mask_Bx] -= 1 # --- Update BOS flag (Original) --- if not bos_over: bos_over = all( dec_step - prefill_step > max_delay_pattern for prefill_step in dec_output.prefill_steps ) dec_output.update_one(pred_BxC, current_step_idx, not bos_over) dec_step += 1 if verbose and dec_step % 86 == 0: duration = time.time() - start_time if duration > 0: print( f"generate step {dec_step}: speed={86 * batch_size / duration:.3f} tokens/s, realtime factor={batch_size / duration:.3f}x" ) start_time = time.time() # --- Finalize and Extract Output --- final_step = dec_step + 1 finished_step_Bx[finished_step_Bx == -1] = final_step - max_delay_pattern prefill_steps_tensor = torch.tensor(dec_output.prefill_steps, device=self.device) lengths_Bx = finished_step_Bx - prefill_steps_tensor lengths_Bx = torch.clamp(lengths_Bx, min=0) max_len = lengths_Bx.max().item() + max_delay_pattern outputs = [] if max_len > 0: num_channels = self.config.decoder_config.num_channels audio_pad_value = self.config.pad_token_id generated_codes = torch.full( (batch_size, max_len, num_channels), fill_value=audio_pad_value, dtype=torch.long, device=self.device, ) for i in range(batch_size): start_step = dec_output.prefill_steps[i] actual_len = lengths_Bx[i].item() + max_delay_pattern if actual_len > 0: tokens_to_copy = dec_output.generated_tokens[i, start_step : start_step + actual_len, :] generated_codes[i, :actual_len, :] = tokens_to_copy if verbose: avg_steps = lengths_Bx.float().mean().item() total_duration = time.time() - total_start_time print(f"generate: avg steps={avg_steps:.1f}, total duration={total_duration:.3f}s") del dec_state outputs = self._generate_output(generated_codes, lengths_Bx) else: print("Warning: Nothing generated for any sequence in the batch.") outputs = [None] * batch_size return outputs if batch_size > 1 else outputs[0] ================================================ FILE: dia/state.py ================================================ from dataclasses import dataclass from typing import Optional import torch from .config import DiaConfig def create_attn_mask( q_padding_mask_1d: torch.Tensor, k_padding_mask_1d: torch.Tensor, device: torch.device, is_causal: bool = False, ) -> torch.Tensor: """ Creates the attention mask (self or cross) mimicking JAX segment ID logic. """ # B1, Tq = q_padding_mask_1d.shape # B2, Tk = k_padding_mask_1d.shape p_mask_q = q_padding_mask_1d.unsqueeze(2) # Shape [B, Tq, 1] p_mask_k = k_padding_mask_1d.unsqueeze(1) # Shape [B, 1, Tk] # Condition A: Non-padding query attends to non-padding key non_pad_attends_non_pad = p_mask_q & p_mask_k # Shape [B, Tq, Tk] # Condition B: Padding query attends to padding key pad_attends_pad = (~p_mask_q) & (~p_mask_k) # Shape [B, Tq, Tk] # Combine: True if padding status is compatible (both non-pad OR both pad) mask = non_pad_attends_non_pad | pad_attends_pad # Shape [B, Tq, Tk] if is_causal: # assert Tq == Tk, "Causal mask requires query and key sequence lengths to be equal" causal_mask_2d = torch.tril(torch.ones_like(mask[0], dtype=torch.bool, device=device)) # Shape [B, Tq, Tk] causal_mask = mask & causal_mask_2d # Shape [B, Tq, Tk] return causal_mask.unsqueeze(1) # Shape [B, 1, Tq, Tk] else: return mask.unsqueeze(1) # Shape [B, 1, Tq, Tk] @dataclass class EncoderInferenceState: """Parameters specifically for encoder inference.""" max_seq_len: int device: torch.device positions: torch.Tensor padding_mask: torch.Tensor attn_mask: torch.Tensor @classmethod def new(cls, config: DiaConfig, cond_src: torch.Tensor) -> "EncoderInferenceState": """Creates EtorchrInferenceParams from DiaConfig and a device.""" device = cond_src.device positions = torch.arange( config.encoder_config.max_position_embeddings, dtype=torch.float32, device=device ).unsqueeze(0) padding_mask = (cond_src.squeeze(1) != 0).to(device).repeat_interleave(2, dim=0) attn_mask = create_attn_mask(padding_mask, padding_mask, device, is_causal=False) return cls( max_seq_len=config.encoder_config.max_position_embeddings, device=device, positions=positions, padding_mask=padding_mask, attn_mask=attn_mask, ) class KVCache(torch.nn.Module): k: torch.Tensor v: torch.Tensor def __init__( self, batch_size: int, num_heads: int, max_len: int, head_dim: int, dtype: torch.dtype, device: torch.device, k: torch.Tensor | None = None, v: torch.Tensor | None = None, ): k = torch.zeros((2 * batch_size, num_heads, max_len, head_dim), dtype=dtype, device=device) if k is None else k v = torch.zeros((2 * batch_size, num_heads, max_len, head_dim), dtype=dtype, device=device) if v is None else v super().__init__() self.register_buffer("k", k) self.register_buffer("v", v) @classmethod def from_kv(cls, k: torch.Tensor, v: torch.Tensor) -> "KVCache": return cls( batch_size=k.shape[0] // 2, num_heads=k.shape[1], max_len=k.shape[2], head_dim=k.shape[3], dtype=k.dtype, device=k.device, k=k, v=v, ) def update(self, k: torch.Tensor, v: torch.Tensor, current_idx: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]: k_out, v_out = self.k, self.v k_out[:, :, current_idx, :] = k v_out[:, :, current_idx, :] = v return self.k, self.v def prefill(self, k: torch.Tensor, v: torch.Tensor): prefill_len = k.shape[2] self.k[:, :, :prefill_len, :] = k self.v[:, :, :prefill_len, :] = v @dataclass class DecoderInferenceState: """Parameters specifically for decoder inference.""" device: torch.device dtype: torch.dtype enc_out: torch.Tensor enc_positions: torch.Tensor dec_positions: torch.Tensor self_attn_cache: list[KVCache] cross_attn_cache: list[KVCache] casual_attn_mask: torch.Tensor cross_attn_mask: torch.Tensor @classmethod def new( cls, config: DiaConfig, enc_state: EncoderInferenceState, enc_out: torch.Tensor, dec_cross_attn_cache: list[KVCache], compute_dtype: torch.dtype, max_generation_length: Optional[int] = None, ) -> "DecoderInferenceState": """Creates DecoderInferenceParams from DiaConfig and a device.""" device = enc_out.device max_audio_len = max_generation_length or config.decoder_config.max_position_embeddings batch_size = enc_out.shape[0] // 2 dec_positions = torch.full((2 * batch_size, 1), fill_value=0, dtype=torch.int32, device=device) causal_mask = torch.tril(torch.ones(max_audio_len, max_audio_len, dtype=torch.bool, device=device)) dec_mask = torch.ones((2 * batch_size, 1), dtype=torch.bool, device=device) cross_attn_mask = create_attn_mask(dec_mask, enc_state.padding_mask, device, is_causal=False) self_attn_cache = [ KVCache( batch_size, config.decoder_config.num_key_value_heads, max_audio_len, config.decoder_config.head_dim, compute_dtype, device, ) for _ in range(config.decoder_config.num_hidden_layers) ] return cls( device=device, dtype=compute_dtype, enc_out=enc_out, enc_positions=enc_state.positions, dec_positions=dec_positions, self_attn_cache=self_attn_cache, cross_attn_cache=dec_cross_attn_cache, casual_attn_mask=causal_mask, cross_attn_mask=cross_attn_mask, ) def prepare_step(self, step_from: int, step_to: int | None = None) -> None: if step_to is None: step_to = step_from + 1 self.dec_positions = torch.arange(step_from, step_to, dtype=torch.int32, device=self.device).unsqueeze(0) @dataclass class DecoderOutput: generated_tokens: torch.Tensor prefill_steps: list[int] @classmethod def new(cls, batch_size: int, config: DiaConfig, device: torch.device) -> "DecoderOutput": max_audio_len = config.decoder_config.max_position_embeddings return cls( generated_tokens=torch.full( (batch_size, max_audio_len, config.decoder_config.num_channels), fill_value=-1, dtype=torch.int, device=device, ), prefill_steps=[], ) def get_tokens_at(self, step_from: int, step_to: int | None = None) -> torch.Tensor: if step_to is None: step_to = step_from + 1 return self.generated_tokens[:, step_from:step_to, :] def update_one(self, dec_out: torch.Tensor, step: int, apply_mask: bool = False): dec_out = dec_out.to(self.generated_tokens.dtype) if apply_mask: mask = self.generated_tokens[:, step, :] == -1 self.generated_tokens[:, step, :] = torch.where(mask, dec_out, self.generated_tokens[:, step, :]) else: self.generated_tokens[:, step, :] = dec_out def prefill(self, dec_out: torch.Tensor, prefill_steps: list[int]): length = dec_out.shape[1] self.generated_tokens[:, :length, :] = dec_out self.prefill_steps = prefill_steps ================================================ FILE: docker/Dockerfile.cpu ================================================ # Dockerfile.cpu - CPU-only deployment for DIA # -------------------------------------------------- # Build: docker build . -f docker/Dockerfile.cpu -t dia-cpu # Run: docker run --rm -p 7860:7860 dia-cpu FROM python:3.10-slim # Set non-interactive frontend ENV DEBIAN_FRONTEND=noninteractive # Install venv, and system dependencies RUN apt-get update && apt-get install -y \ python3-venv \ libsndfile1 \ ffmpeg \ curl \ && apt-get clean && rm -rf /var/lib/apt/lists/* # Create non-root user and set up directories RUN useradd -m -u 1001 appuser && \ mkdir -p /app/outputs /app && \ chown -R appuser:appuser /app USER appuser WORKDIR /app # Copy all code (including pyproject.toml) COPY --chown=appuser:appuser . . # Create and activate virtual environment RUN python3 -m venv /app/venv ENV PATH="/app/venv/bin:$PATH" # Install all project dependencies (CPU-only PyTorch) RUN pip install --upgrade pip && \ pip install torch torchaudio --index-url https://download.pytorch.org/whl/cpu && \ pip install --no-cache-dir -e .[dev] # Set environment variables ENV PYTHONUNBUFFERED=1 \ PYTHONPATH=/app # Expose Gradio default port ENV GRADIO_SERVER_NAME="0.0.0.0" EXPOSE 7860 # Entrypoint CMD ["python3", "app.py"] ================================================ FILE: docker/Dockerfile.gpu ================================================ # Dockerfile.gpu - GPU deployment for DIA # -------------------------------------------------- # Build: docker build . -f docker/Dockerfile.gpu -t dia-gpu # Run: docker run --rm --gpus all -p 7860:7860 dia-gpu # Requires NVIDIA Container Toolkit on host. FROM pytorch/pytorch:2.1.2-cuda12.1-cudnn8-runtime # Set non-interactive frontend ENV DEBIAN_FRONTEND=noninteractive # Install venv, and system dependencies RUN apt-get update && apt-get install -y \ python3-venv \ libsndfile1 \ ffmpeg \ curl \ && apt-get clean && rm -rf /var/lib/apt/lists/* # Create non-root user and set up directories RUN useradd -m -u 1001 appuser && \ mkdir -p /app/outputs /app && \ chown -R appuser:appuser /app USER appuser WORKDIR /app # Copy all code (including pyproject.toml) COPY --chown=appuser:appuser . . # Create and activate virtual environment RUN python3 -m venv /app/venv ENV PATH="/app/venv/bin:$PATH" # Install all project dependencies RUN pip install --upgrade pip && pip install --no-cache-dir . # Set environment variables ENV PYTHONUNBUFFERED=1 \ PYTHONPATH=/app \ USE_GPU=true \ LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/local/cuda-12.1/lib64:${LD_LIBRARY_PATH} # Expose Gradio default port ENV GRADIO_SERVER_NAME="0.0.0.0" EXPOSE 7860 # Entrypoint CMD ["python3", "app.py"] ================================================ FILE: example/benchmark.py ================================================ from random import choice import torch from dia.model import Dia torch._inductor.config.coordinate_descent_tuning = True torch._inductor.config.triton.unique_kernel_names = True torch._inductor.config.fx_graph_cache = True # debugging torch._logging.set_logs(graph_breaks=True, recompiles=True) model_name = "nari-labs/Dia-1.6B-0626" compute_dtype = "float16" model = Dia.from_pretrained(model_name, compute_dtype=compute_dtype) test_cases = [ "[S1] Dia is an open weights text to dialogue model.", "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face.", "[S1] torch.compile is a new feature in PyTorch that allows you to compile your model with a single line of code.", "[S1] torch.compile is a new feature in PyTorch that allows you to compile your model with a single line of code. [S2] It is a new feature in PyTorch that allows you to compile your model with a single line of code.", ] # Wram up for _ in range(2): text = choice(test_cases) output = model.generate(text, audio_prompt="./example_prompt.mp3", use_torch_compile=True, verbose=True) output = model.generate(text, use_torch_compile=True, verbose=True) # Benchmark for _ in range(10): text = choice(test_cases) output = model.generate(text, use_torch_compile=True, verbose=True) output = model.generate(text, audio_prompt="./example_prompt.mp3", use_torch_compile=True, verbose=True) ================================================ FILE: example/simple-cpu.py ================================================ import torch from dia.model import Dia # Select device: CPU device = torch.device("cpu") print(f"Using device: {device}") # Load model model = Dia.from_pretrained( "nari-labs/Dia-1.6B-0626", compute_dtype="float32", device=device ) # Float32 works better than float16 on CPU - you can also test with float16 text = "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." output = model.generate(text, use_torch_compile=False, verbose=True) model.save_audio("simple.mp3", output) ================================================ FILE: example/simple-mac.py ================================================ from dia.model import Dia model = Dia.from_pretrained("nari-labs/Dia-1.6B-0626", compute_dtype="float16") text = "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." # It is important to set the `use_torch_compile` argument to `False` when using Dia on MacOS. # This is because the `torch.compile` function is not supported on MacOS. output = model.generate(text, use_torch_compile=False, verbose=True) model.save_audio("simple.mp3", output) ================================================ FILE: example/simple.py ================================================ from dia.model import Dia model = Dia.from_pretrained("nari-labs/Dia-1.6B-0626", compute_dtype="float16") text = "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." output = model.generate( text, use_torch_compile=False, verbose=True, cfg_scale=3.0, temperature=1.8, top_p=0.90, cfg_filter_top_k=50, ) model.save_audio("simple.mp3", output) ================================================ FILE: example/simple_batch.py ================================================ from dia.model import Dia model = Dia.from_pretrained("nari-labs/Dia-1.6B-0626", compute_dtype="float16") text = "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." texts = [text for _ in range(10)] output = model.generate(texts, use_torch_compile=True, verbose=True, max_tokens=1500) for i, o in enumerate(output): model.save_audio(f"simple_{i}.mp3", o) ================================================ FILE: example/voice_clone.py ================================================ from dia.model import Dia model = Dia.from_pretrained("nari-labs/Dia-1.6B-0626", compute_dtype="float16") # You should put the transcript of the voice you want to clone # We will use the audio created by running simple.py as an example. # Note that you will be REQUIRED TO RUN simple.py for the script to work as-is. clone_from_text = "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." clone_from_audio = "simple.mp3" # For your custom needs, replace above with below and add your audio file to this directory: # clone_from_text = "[S1] ... [S2] ... [S1] ... corresponding to your_audio_name.mp3" # clone_from_audio = "your_audio_name.mp3" # Text to generate text_to_generate = "[S1] Hello, how are you? [S2] I'm good, thank you. [S1] What's your name? [S2] My name is Dia. [S1] Nice to meet you. [S2] Nice to meet you too." # It will only return the audio from the text_to_generate output = model.generate( clone_from_text + text_to_generate, audio_prompt=clone_from_audio, use_torch_compile=False, verbose=True, cfg_scale=4.0, temperature=1.8, top_p=0.90, cfg_filter_top_k=50, ) model.save_audio("voice_clone.mp3", output) ================================================ FILE: example/voice_clone_batch.py ================================================ from dia.model import Dia model = Dia.from_pretrained("nari-labs/Dia-1.6B-0626", compute_dtype="float16") # You should put the transcript of the voice you want to clone # We will use the audio created by running simple.py as an example. # Note that you will be REQUIRED TO RUN simple.py for the script to work as-is. clone_from_text = "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." # For your custom needs, replace above with below and add your audio file to this directory: # clone_from_text = "[S1] ... [S2] ... [S1] ... corresponding to your_audio_name.mp3" # clone_from_audio = "your_audio_name.mp3" # Text to generate text_to_generate = "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." clone_from_audios = [f"simple_{i}.mp3" for i in range(10)] texts = [clone_from_text + text_to_generate for _ in range(10)] # It will only return the audio from the text_to_generate output = model.generate(texts, audio_prompt=clone_from_audios, use_torch_compile=True, verbose=True, max_tokens=2000) for i, o in enumerate(output): model.save_audio(f"voice_clone_{i}.mp3", o) ================================================ FILE: hf.py ================================================ from transformers import AutoProcessor, DiaForConditionalGeneration torch_device = "cuda" model_checkpoint = "nari-labs/Dia-1.6B-0626" text = [ "[S1] Dia is an open weights text to dialogue model. [S2] You get full control over scripts and voices. [S1] Wow. Amazing. (laughs) [S2] Try it now on Git hub or Hugging Face." ] processor = AutoProcessor.from_pretrained(model_checkpoint) inputs = processor(text=text, padding=True, return_tensors="pt").to(torch_device) model = DiaForConditionalGeneration.from_pretrained(model_checkpoint).to(torch_device) outputs = model.generate(**inputs, max_new_tokens=3072, guidance_scale=3.0, temperature=1.8, top_p=0.90, top_k=45) outputs = processor.batch_decode(outputs) processor.save_audio(outputs, "example.mp3") ================================================ FILE: pyproject.toml ================================================ [project] name = "nari-tts" version = "0.1.0" description = "Dia - A text-to-speech model for dialogue generation" readme = "README.md" requires-python = ">=3.10" license = {file = "LICENSE"} authors = [ {name = "Nari Labs", email = "contact@narilabs.ai"} ] dependencies = [ "descript-audio-codec>=1.0.0", "gradio>=5.25.2", "huggingface-hub>=0.30.2", "numpy>=2.2.4", "pydantic>=2.11.3", "safetensors>=0.5.3", "soundfile>=0.13.1", "torch==2.6.0", "torchaudio==2.6.0", "triton==3.2.0 ; sys_platform == 'linux'", "triton-windows==3.2.0.post18 ; sys_platform == 'win32'", ] [build-system] requires = ["hatchling"] build-backend = "hatchling.build" [project.urls] "Homepage" = "https://github.com/nari-labs/dia" "Bug Tracker" = "https://github.com/nari-labs/dia/issues" [tool.hatch.build.targets.wheel] packages = ["dia"] [tool.ruff] # Never enforce `E501` (line length violations). lint.ignore = ["C901", "E501", "E741", "W605"] lint.select = ["C", "E", "F", "I", "W"] line-length = 119 # Ignore import violations in all `__init__.py` files. [tool.ruff.lint.per-file-ignores] "__init__.py" = ["E402", "F401", "F403", "F811"] [tool.ruff.lint.isort] lines-after-imports = 2 [tool.uv.sources] torch = [ { index = "pytorch-cu126", marker = "sys_platform == 'linux' or sys_platform == 'win32'" }, ] torchaudio = [ { index = "pytorch-cu126", marker = "sys_platform == 'linux' or sys_platform == 'win32'" }, ] [[tool.uv.index]] name = "pytorch-cu126" url = "https://download.pytorch.org/whl/cu126" explicit = true [dependency-groups] dev = [ "ninja>=1.11.1.4", "packaging>=25.0", ]