Repository: Lightricks/LTX-Video
Branch: main
Commit: 4b2d05305762
Files: 54
Total size: 418.1 KB
Directory structure:
gitextract_jt2joi_e/
├── .gitattributes
├── .github/
│ └── workflows/
│ └── pylint.yml
├── .gitignore
├── .pre-commit-config.yaml
├── LICENSE
├── README.md
├── configs/
│ ├── ltxv-13b-0.9.8-dev-fp8.yaml
│ ├── ltxv-13b-0.9.8-dev.yaml
│ ├── ltxv-13b-0.9.8-distilled-fp8.yaml
│ ├── ltxv-13b-0.9.8-distilled.yaml
│ ├── ltxv-2b-0.9.1.yaml
│ ├── ltxv-2b-0.9.5.yaml
│ ├── ltxv-2b-0.9.6-dev.yaml
│ ├── ltxv-2b-0.9.6-distilled.yaml
│ ├── ltxv-2b-0.9.8-distilled-fp8.yaml
│ ├── ltxv-2b-0.9.8-distilled.yaml
│ └── ltxv-2b-0.9.yaml
├── inference.py
├── ltx_video/
│ ├── __init__.py
│ ├── inference.py
│ ├── models/
│ │ ├── __init__.py
│ │ ├── autoencoders/
│ │ │ ├── __init__.py
│ │ │ ├── causal_conv3d.py
│ │ │ ├── causal_video_autoencoder.py
│ │ │ ├── conv_nd_factory.py
│ │ │ ├── dual_conv3d.py
│ │ │ ├── latent_upsampler.py
│ │ │ ├── pixel_norm.py
│ │ │ ├── pixel_shuffle.py
│ │ │ ├── vae.py
│ │ │ ├── vae_encode.py
│ │ │ └── video_autoencoder.py
│ │ └── transformers/
│ │ ├── __init__.py
│ │ ├── attention.py
│ │ ├── embeddings.py
│ │ ├── symmetric_patchifier.py
│ │ └── transformer3d.py
│ ├── pipelines/
│ │ ├── __init__.py
│ │ ├── crf_compressor.py
│ │ └── pipeline_ltx_video.py
│ ├── schedulers/
│ │ ├── __init__.py
│ │ └── rf.py
│ └── utils/
│ ├── __init__.py
│ ├── diffusers_config_mapping.py
│ ├── prompt_enhance_utils.py
│ ├── skip_layer_strategy.py
│ └── torch_utils.py
├── pyproject.toml
└── tests/
├── conftest.py
├── test_configs.py
├── test_inference.py
├── test_scheduler.py
├── test_vae.py
└── utils/
└── .gitattributes
================================================
FILE CONTENTS
================================================
================================================
FILE: .gitattributes
================================================
*.jpg filter=lfs diff=lfs merge=lfs -text
*.jpeg filter=lfs diff=lfs merge=lfs -text
*.png filter=lfs diff=lfs merge=lfs -text
*.gif filter=lfs diff=lfs merge=lfs -text
tests/utils/car.png filter=lfs diff=lfs merge=lfs -text
================================================
FILE: .github/workflows/pylint.yml
================================================
name: Ruff
on: [push]
jobs:
build:
runs-on: ubuntu-latest
strategy:
matrix:
python-version: ["3.10"]
steps:
- name: Checkout repository and submodules
uses: actions/checkout@v4
- name: Set up Python ${{ matrix.python-version }}
uses: actions/setup-python@v5
with:
python-version: ${{ matrix.python-version }}
- name: Install dependencies
run: |
python -m pip install --upgrade pip
pip install ruff==0.2.2 black==24.2.0
- name: Analyzing the code with ruff
run: |
ruff $(git ls-files '*.py')
- name: Verify that no Black changes are required
run: |
black --check $(git ls-files '*.py')
================================================
FILE: .gitignore
================================================
# Byte-compiled / optimized / DLL files
__pycache__/
*.py[cod]
*$py.class
# C extensions
*.so
# Distribution / packaging
.Python
build/
develop-eggs/
dist/
downloads/
eggs/
.eggs/
lib/
lib64/
parts/
sdist/
var/
wheels/
share/python-wheels/
*.egg-info/
.installed.cfg
*.egg
MANIFEST
# PyInstaller
# Usually these files are written by a python script from a template
# before PyInstaller builds the exe, so as to inject date/other infos into it.
*.manifest
*.spec
# Installer logs
pip-log.txt
pip-delete-this-directory.txt
# Unit test / coverage reports
htmlcov/
.tox/
.nox/
.coverage
.coverage.*
.cache
nosetests.xml
coverage.xml
*.cover
*.py,cover
.hypothesis/
.pytest_cache/
cover/
# Translations
*.mo
*.pot
# Django stuff:
*.log
local_settings.py
db.sqlite3
db.sqlite3-journal
# Flask stuff:
instance/
.webassets-cache
# Scrapy stuff:
.scrapy
# Sphinx documentation
docs/_build/
# PyBuilder
.pybuilder/
target/
# Jupyter Notebook
.ipynb_checkpoints
# IPython
profile_default/
ipython_config.py
# pyenv
# For a library or package, you might want to ignore these files since the code is
# intended to run in multiple environments; otherwise, check them in:
# .python-version
# pipenv
# According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
# However, in case of collaboration, if having platform-specific dependencies or dependencies
# having no cross-platform support, pipenv may install dependencies that don't work, or not
# install all needed dependencies.
#Pipfile.lock
# poetry
# Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
# This is especially recommended for binary packages to ensure reproducibility, and is more
# commonly ignored for libraries.
# https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
#poetry.lock
# pdm
# Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
#pdm.lock
# pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
# in version control.
# https://pdm.fming.dev/latest/usage/project/#working-with-version-control
.pdm.toml
.pdm-python
.pdm-build/
# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
__pypackages__/
# Celery stuff
celerybeat-schedule
celerybeat.pid
# SageMath parsed files
*.sage.py
# Environments
.env
.venv
env/
venv/
ENV/
env.bak/
venv.bak/
# Spyder project settings
.spyderproject
.spyproject
# Rope project settings
.ropeproject
# mkdocs documentation
/site
# mypy
.mypy_cache/
.dmypy.json
dmypy.json
# Pyre type checker
.pyre/
# pytype static type analyzer
.pytype/
# Cython debug symbols
cython_debug/
# PyCharm
# JetBrains specific template is maintained in a separate JetBrains.gitignore that can
# be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
# and can be added to the global gitignore or merged into this file. For a more nuclear
# option (not recommended) you can uncomment the following to ignore the entire idea folder.
.idea/
# From inference.py
outputs/
*.mp4
*.png
!tests/utils/car.png
================================================
FILE: .pre-commit-config.yaml
================================================
repos:
- repo: https://github.com/astral-sh/ruff-pre-commit
# Ruff version.
rev: v0.2.2
hooks:
# Run the linter.
- id: ruff
args: [--fix] # Automatically fix issues if possible.
types: [python] # Ensure it only runs on .py files.
- repo: https://github.com/psf/black
rev: 24.2.0 # Specify the version of Black you want
hooks:
- id: black
name: Black code formatter
language_version: python3 # Use the Python version you're targeting (e.g., 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 [yyyy] [name of copyright owner]
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
================================================
# LTX-Video
[](https://ltx.video)
[](https://huggingface.co/Lightricks/LTX-Video)
[](https://app.ltx.studio/ltx-2-playground/t2v)
[](https://arxiv.org/abs/2501.00103)
[](https://github.com/Lightricks/LTX-Video-Trainer)
[](https://discord.gg/ltxplatform)
This is the official repository for LTX-Video.
---
## 🚀 **New: LTX-2 is Now Available!**
**We're excited to announce [LTX-2](https://github.com/Lightricks/LTX-2) - the next generation of LTX with synchronized audio+video generation!**
LTX-2 is the first DiT-based audio-video foundation model that contains all core capabilities of modern video generation in one model. **LTX-2 is now the primary home for LTX development** and includes significant improvements:
- 🎵 **Synchronized Audio+Video Generation** - Generate videos with perfectly synchronized audio
- 🎬 **Latest Model** - LTX-2 with improved quality and capabilities
- 🔌 **ComfyUI Integration** - Built into ComfyUI core for seamless workflows
- 🎯 **Advanced Features:**
- Multiple keyframe support
- IC-LoRA control models for precise generation
- Standard LoRA support for style customization
- Latent upsampler for multiscale pipelines
- 🛠️ **Training Tools** - LoRA training capabilities
- 📚 **Comprehensive Documentation** - Full documentation at [https://docs.ltx.video](https://docs.ltx.video)
- 🔄 **Active Development** - Ongoing improvements and community support
**[👉 Check out LTX-2 here](https://github.com/Lightricks/LTX-2)**
**[📖 View Documentation](https://docs.ltx.video)**
---
## Table of Contents
- [Introduction](#introduction)
- [What's New](#news)
- [Models](#models)
- [Quick Start Guide](#quick-start-guide)
- [Online demo](#online-inference)
- [Run locally](#run-locally)
- [Installation](#installation)
- [Inference](#inference)
- [ComfyUI Integration](#comfyui-integration)
- [Diffusers Integration](#diffusers-integration)
- [Model User Guide](#model-user-guide)
- [Community Contribution](#community-contribution)
- [Training](#training)
- [Control Models](#control-models)
- [Join Us!](#join-us)
- [Acknowledgement](#acknowledgement)
# Introduction
LTX-Video is the first DiT-based video generation model that contains all core capabilities of modern video generation in one model: synchronized audio and video, high fidelity, multiple performance modes, production-ready outputs, API access, and open access. It can generate up to 50 FPS videos at native 4K resolution with synchronized audio in one pass.
The model is trained on a large-scale dataset of diverse videos and can generate high-resolution videos with realistic and diverse content.
The model supports image-to-video, multi-keyframe conditioning, keyframe-based animation, video extension (both forward and backward), video-to-video transformations, and any combination of these features.
### Image-to-video examples
| | | |
|:---:|:---:|:---:|
|  |  |  |
|  |  |  |
|  |  |  |
### Controlled video examples
| | | |
|:---:|:---:|:---:|
|  |  |  |
| | |
|:---:|:---:|
|  |  |
# News
## October 23, 2025: LTX-2 Announced
Today we announced our newest foundation model, LTX-2. LTX-2 represents a major leap forward from our previous model, LTXV 0.9.8. Here’s what’s new:
* **Audio + Video, Together**: Visuals and sound are generated in one coherent process, with motion, dialogue, ambience, and music flowing simultaneously.
* **4K Fidelity**: Professional-grade precision with native 4K and up to 50 fps, sharp textures, clean motion, and synchronized audio.
* **Longer Generations**: LTX-2 supports longer, continuous clips with synchronized audio up to 10 seconds.
* **Low Cost & Efficiency**: Up to 50% lower compute cost than competing models, powered by a multi-GPU inference stack.
* **Creative Control**: Multi-keyframe conditioning, 3D camera logic, and LoRA fine-tuning deliver frame-level precision and style consistency.
For more details, please see our [blog post](https://website.ltx.video/blog/introducing-ltx-2). LTX-2 model weights, code, and benchmarks will be released to the community later in 2025.
## July, 16th, 2025: New Distilled models v0.9.8 with up to 60 seconds of video:
- Long shot generation in LTXV-13B!
* LTX-Video now supports up to 60 seconds of video.
* Compatible also with the official IC-LoRAs.
* Try now in [ComfyUI](https://github.com/Lightricks/ComfyUI-LTXVideo/tree/master/example_workflows/ltxv-13b-i2v-long-multi-prompt.json).
- Release a new distilled models:
* 13B distilled model [ltxv-13b-0.9.8-distilled](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-13b-0.9.8-distilled.yaml)
* 2B distilled model [ltxv-2b-0.9.8-distilled](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-2b-0.9.8-distilled.yaml)
* Both models are distilled from the same base model [ltxv-13b-0.9.8-dev](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-13b-0.9.8-dev.yaml) and are compatible for use together in the same multiscale pipeline.
* Improved prompt understanding and detail generation
* Includes corresponding FP8 weights and workflows.
- Release a new detailer model [LTX-Video-ICLoRA-detailer-13B-0.9.8](https://huggingface.co/Lightricks/LTX-Video-ICLoRA-detailer-13b-0.9.8)
* Available in [ComfyUI](https://github.com/Lightricks/ComfyUI-LTXVideo/tree/master/example_workflows/ltxv-13b-upscale.json).
## July, 8th, 2025: New Control Models Released!
- Released three new control models for LTX-Video on HuggingFace:
* **Depth Control**: [LTX-Video-ICLoRA-depth-13b-0.9.7](https://huggingface.co/Lightricks/LTX-Video-ICLoRA-depth-13b-0.9.7)
* **Pose Control**: [LTX-Video-ICLoRA-pose-13b-0.9.7](https://huggingface.co/Lightricks/LTX-Video-ICLoRA-pose-13b-0.9.7)
* **Canny Control**: [LTX-Video-ICLoRA-canny-13b-0.9.7](https://huggingface.co/Lightricks/LTX-Video-ICLoRA-canny-13b-0.9.7)
## May, 14th, 2025: New distilled model 13B v0.9.7:
- Release a new 13B distilled model [ltxv-13b-0.9.7-distilled](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-13b-0.9.7-distilled.safetensors)
* Amazing for iterative work - generates HD videos in 10 seconds, with low-res preview after just 3 seconds (on H100)!
* Does not require classifier-free guidance and spatio-temporal guidance.
* Supports sampling with 8 (recommended), or less diffusion steps.
* Also released a LoRA version of the distilled model, [ltxv-13b-0.9.7-distilled-lora128](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-13b-0.9.7-distilled-lora128.safetensors)
* Requires only 1GB of VRAM
* Can be used with the full 13B model for fast inference
- Release a new quantized distilled model [ltxv-13b-0.9.7-distilled-fp8](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-13b-0.9.7-distilled-fp8.safetensors) for *real-time* generation (on H100) with even less VRAM
## May, 5th, 2025: New model 13B v0.9.7:
- Release a new 13B model [ltxv-13b-0.9.7-dev](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-13b-0.9.7-dev.safetensors)
- Release a new quantized model [ltxv-13b-0.9.7-dev-fp8](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-13b-0.9.7-dev-fp8.safetensors) for faster inference with less VRam
- Release a new upscalers
* [ltxv-temporal-upscaler-0.9.7](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-temporal-upscaler-0.9.7.safetensors)
* [ltxv-spatial-upscaler-0.9.7](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-spatial-upscaler-0.9.7.safetensors)
- Breakthrough prompt adherence and physical understanding.
- New Pipeline for multi-scale video rendering for fast and high quality results
## April, 15th, 2025: New checkpoints v0.9.6:
- Release a new checkpoint [ltxv-2b-0.9.6-dev-04-25](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-2b-0.9.6-dev-04-25.safetensors) with improved quality
- Release a new distilled model [ltxv-2b-0.9.6-distilled-04-25](https://huggingface.co/Lightricks/LTX-Video/blob/main/ltxv-2b-0.9.6-distilled-04-25.safetensors)
* 15x faster inference than non-distilled model.
* Does not require classifier-free guidance and spatio-temporal guidance.
* Supports sampling with 8 (recommended), or less diffusion steps.
- Improved prompt adherence, motion quality and fine details.
- New default resolution and FPS: 1216 × 704 pixels at 30 FPS
* Still real time on H100 with the distilled model.
* Other resolutions and FPS are still supported.
- Support stochastic inference (can improve visual quality when using the distilled model)
## March, 5th, 2025: New checkpoint v0.9.5
- New license for commercial use ([OpenRail-M](https://huggingface.co/Lightricks/LTX-Video/ltx-video-2b-v0.9.5.license.txt))
- Release a new checkpoint v0.9.5 with improved quality
- Support keyframes and video extension
- Support higher resolutions
- Improved prompt understanding
- Improved VAE
- New online web app in [LTX-Studio](https://app.ltx.studio/ltx-video)
- Automatic prompt enhancement
## February, 20th, 2025: More inference options
- Improve STG (Spatiotemporal Guidance) for LTX-Video
- Support MPS on macOS with PyTorch 2.3.0
- Add support for 8-bit model, LTX-VideoQ8
- Add TeaCache for LTX-Video
- Add [ComfyUI-LTXTricks](#comfyui-integration)
- Add Diffusion-Pipe
## December 31st, 2024: Research paper
- Release the [research paper](https://arxiv.org/abs/2501.00103)
## December 20th, 2024: New checkpoint v0.9.1
- Release a new checkpoint v0.9.1 with improved quality
- Support for STG / PAG
- Support loading checkpoints of LTX-Video in Diffusers format (conversion is done on-the-fly)
- Support offloading unused parts to CPU
- Support the new timestep-conditioned VAE decoder
- Reference contributions from the community in the readme file
- Relax transformers dependency
## November 21th, 2024: Initial release v0.9.0
- Initial release of LTX-Video
- Support text-to-video and image-to-video generation
# Models
| Name | Notes | inference.py config | ComfyUI workflow (Recommended) |
|-------------------------|--------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------------------------------------------------------|--------------------------------|
| ltxv-13b-0.9.8-dev | Highest quality, requires more VRAM | [ltxv-13b-0.9.8-dev.yaml](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-13b-0.9.8-dev.yaml) | [ltxv-13b-i2v-base.json](https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/ltxv-13b-i2v-base.json) |
| [ltxv-13b-0.9.8-mix](https://app.ltx.studio/motion-workspace?videoModel=ltxv-13b) | Mix ltxv-13b-dev and ltxv-13b-distilled in the same multi-scale rendering workflow for balanced speed-quality | N/A | [ltxv-13b-i2v-mixed-multiscale.json](https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/ltxv-13b-i2v-mixed-multiscale.json) |
[ltxv-13b-0.9.8-distilled](https://app.ltx.studio/motion-workspace?videoModel=ltxv) | Faster, less VRAM usage, slight quality reduction compared to 13b. Ideal for rapid iterations | [ltxv-13b-0.9.8-distilled.yaml](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-13b-0.9.8-distilled.yaml) | [ltxv-13b-dist-i2v-base.json](https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/13b-distilled/ltxv-13b-dist-i2v-base.json) |
ltxv-2b-0.9.8-distilled | Smaller model, slight quality reduction compared to 13b distilled. Ideal for fast generation with light VRAM usage | [ltxv-2b-0.9.8-distilled.yaml](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-2b-0.9.8-distilled.yaml) | N/A |
| ltxv-13b-0.9.8-dev-fp8 | Quantized version of ltxv-13b | [ltxv-13b-0.9.8-dev-fp8.yaml](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-13b-0.9.8-dev-fp8.yaml) | [ltxv-13b-i2v-base-fp8.json](https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/ltxv-13b-i2v-base-fp8.json) |
| ltxv-13b-0.9.8-distilled-fp8 | Quantized version of ltxv-13b-distilled | [ltxv-13b-0.9.8-distilled-fp8.yaml](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-13b-0.9.8-distilled-fp8.yaml) | [ltxv-13b-dist-i2v-base-fp8.json](https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/13b-distilled/ltxv-13b-dist-i2v-base-fp8.json) |
| ltxv-2b-0.9.8-distilled-fp8 | Quantized version of ltxv-2b-distilled | [ltxv-2b-0.9.8-distilled-fp8.yaml](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-2b-0.9.8-distilled-fp8.yaml) | N/A |
| ltxv-2b-0.9.6 | Good quality, lower VRAM requirement than ltxv-13b | [ltxv-2b-0.9.6-dev.yaml](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-2b-0.9.6-dev.yaml) | [ltxvideo-i2v.json](https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/low_level/ltxvideo-i2v.json) |
| ltxv-2b-0.9.6-distilled | 15× faster, real-time capable, fewer steps needed, no STG/CFG required | [ltxv-2b-0.9.6-distilled.yaml](https://github.com/Lightricks/LTX-Video/blob/main/configs/ltxv-2b-0.9.6-distilled.yaml) | [ltxvideo-i2v-distilled.json](https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/low_level/ltxvideo-i2v-distilled.json) |
# Quick Start Guide
## Online inference
The model is accessible right away via the following links:
- [LTX-Studio image-to-video (13B-mix)](https://app.ltx.studio/motion-workspace?videoModel=ltxv-13b)
- [LTX-Studio image-to-video (13B distilled)](https://app.ltx.studio/motion-workspace?videoModel=ltxv)
- [Fal.ai image-to-video (13B full)](https://fal.ai/models/fal-ai/ltx-video-13b-dev/image-to-video)
- [Fal.ai image-to-video (13B distilled)](https://fal.ai/models/fal-ai/ltx-video-13b-distilled/image-to-video)
- [Replicate image-to-video](https://replicate.com/lightricks/ltx-video)
## Run locally
### Installation
The codebase was tested with Python 3.10.5, CUDA version 12.2, and supports PyTorch >= 2.1.2.
On macOS, MPS was tested with PyTorch 2.3.0, and should support PyTorch == 2.3 or >= 2.6.
```bash
git clone https://github.com/Lightricks/LTX-Video.git
cd LTX-Video
# create env
python -m venv env
source env/bin/activate
python -m pip install -e .\[inference\]
```
#### FP8 Kernels (optional)
[FP8 kernels](https://github.com/Lightricks/LTXVideo-Q8-Kernels) developed for LTX-Video provide performance boost on supported graphics cards (Ada architecture and later). To install FP8 kernels, follow the instructions in that repository.
### Inference
📝 **Note:** For best results, we recommend using our [ComfyUI](#comfyui-integration) workflow. We're working on updating the inference.py script to match the high quality and output fidelity of ComfyUI.
To use our model, please follow the inference code in [inference.py](./inference.py):
#### For image-to-video generation:
```bash
python inference.py --prompt "PROMPT" --conditioning_media_paths IMAGE_PATH --conditioning_start_frames 0 --height HEIGHT --width WIDTH --num_frames NUM_FRAMES --seed SEED --pipeline_config configs/ltxv-13b-0.9.8-distilled.yaml
```
#### Extending a video:
📝 **Note:** Input video segments must contain a multiple of 8 frames plus 1 (e.g., 9, 17, 25, etc.), and the target frame number should be a multiple of 8.
```bash
python inference.py --prompt "PROMPT" --conditioning_media_paths VIDEO_PATH --conditioning_start_frames START_FRAME --height HEIGHT --width WIDTH --num_frames NUM_FRAMES --seed SEED --pipeline_config configs/ltxv-13b-0.9.8-distilled.yaml
```
#### For video generation with multiple conditions:
You can now generate a video conditioned on a set of images and/or short video segments.
Simply provide a list of paths to the images or video segments you want to condition on, along with their target frame numbers in the generated video. You can also specify the conditioning strength for each item (default: 1.0).
```bash
python inference.py --prompt "PROMPT" --conditioning_media_paths IMAGE_OR_VIDEO_PATH_1 IMAGE_OR_VIDEO_PATH_2 --conditioning_start_frames TARGET_FRAME_1 TARGET_FRAME_2 --height HEIGHT --width WIDTH --num_frames NUM_FRAMES --seed SEED --pipeline_config configs/ltxv-13b-0.9.8-distilled.yaml
```
### Using as a library
```python
from ltx_video.inference import infer, InferenceConfig
infer(
InferenceConfig(
pipeline_config="configs/ltxv-13b-0.9.8-distilled.yaml",
prompt=PROMPT,
height=HEIGHT,
width=WIDTH,
num_frames=NUM_FRAMES,
output_path="output.mp4",
)
)
```
## ComfyUI Integration
To use our model with ComfyUI, please follow the instructions at [https://github.com/Lightricks/ComfyUI-LTXVideo/](https://github.com/Lightricks/ComfyUI-LTXVideo/).
## Diffusers Integration
To use our model with the Diffusers Python library, check out the [official documentation](https://huggingface.co/docs/diffusers/main/en/api/pipelines/ltx_video).
Diffusers also support an 8-bit version of LTX-Video, [see details below](#ltx-videoq8)
# Model User Guide
## 📝 Prompt Engineering
When writing prompts, focus on detailed, chronological descriptions of actions and scenes. Include specific movements, appearances, camera angles, and environmental details - all in a single flowing paragraph. Start directly with the action, and keep descriptions literal and precise. Think like a cinematographer describing a shot list. Keep within 200 words. For best results, build your prompts using this structure:
* Start with main action in a single sentence
* Add specific details about movements and gestures
* Describe character/object appearances precisely
* Include background and environment details
* Specify camera angles and movements
* Describe lighting and colors
* Note any changes or sudden events
* See [examples](#introduction) for more inspiration.
### Automatic Prompt Enhancement
When using `LTXVideoPipeline` directly, you can enable prompt enhancement by setting `enhance_prompt=True`.
## 🎮 Parameter Guide
* Resolution Preset: Higher resolutions for detailed scenes, lower for faster generation and simpler scenes. The model works on resolutions that are divisible by 32 and number of frames that are divisible by 8 + 1 (e.g. 257). In case the resolution or number of frames are not divisible by 32 or 8 + 1, the input will be padded with -1 and then cropped to the desired resolution and number of frames. The model works best on resolutions under 720 x 1280 and number of frames below 257
* Seed: Save seed values to recreate specific styles or compositions you like
* Guidance Scale: 3-3.5 are the recommended values
* Inference Steps: More steps (40+) for quality, fewer steps (20-30) for speed
📝 For advanced parameters usage, please see `python inference.py --help`
## Community Contribution
### ComfyUI-LTXTricks 🛠️
A community project providing additional nodes for enhanced control over the LTX Video model. It includes implementations of advanced techniques like RF-Inversion, RF-Edit, FlowEdit, and more. These nodes enable workflows such as Image and Video to Video (I+V2V), enhanced sampling via Spatiotemporal Skip Guidance (STG), and interpolation with precise frame settings.
- **Repository:** [ComfyUI-LTXTricks](https://github.com/logtd/ComfyUI-LTXTricks)
- **Features:**
- 🔄 **RF-Inversion:** Implements [RF-Inversion](https://rf-inversion.github.io/) with an [example workflow here](https://github.com/logtd/ComfyUI-LTXTricks/blob/main/example_workflows/example_ltx_inversion.json).
- ✂️ **RF-Edit:** Implements [RF-Solver-Edit](https://github.com/wangjiangshan0725/RF-Solver-Edit) with an [example workflow here](https://github.com/logtd/ComfyUI-LTXTricks/blob/main/example_workflows/example_ltx_rf_edit.json).
- 🌊 **FlowEdit:** Implements [FlowEdit](https://github.com/fallenshock/FlowEdit) with an [example workflow here](https://github.com/logtd/ComfyUI-LTXTricks/blob/main/example_workflows/example_ltx_flow_edit.json).
- 🎥 **I+V2V:** Enables Video to Video with a reference image. [Example workflow](https://github.com/logtd/ComfyUI-LTXTricks/blob/main/example_workflows/example_ltx_iv2v.json).
- ✨ **Enhance:** Partial implementation of [STGuidance](https://junhahyung.github.io/STGuidance/). [Example workflow](https://github.com/logtd/ComfyUI-LTXTricks/blob/main/example_workflows/example_ltxv_stg.json).
- 🖼️ **Interpolation and Frame Setting:** Nodes for precise control of latents per frame. [Example workflow](https://github.com/logtd/ComfyUI-LTXTricks/blob/main/example_workflows/example_ltx_interpolation.json).
### LTX-VideoQ8 🎱
**LTX-VideoQ8** is an 8-bit optimized version of [LTX-Video](https://github.com/Lightricks/LTX-Video), designed for faster performance on NVIDIA ADA GPUs.
- **Repository:** [LTX-VideoQ8](https://github.com/KONAKONA666/LTX-Video)
- **Features:**
- 🚀 Up to 3X speed-up with no accuracy loss
- 🎥 Generate 720x480x121 videos in under a minute on RTX 4060 (8GB VRAM)
- 🛠️ Fine-tune 2B transformer models with precalculated latents
- **Community Discussion:** [Reddit Thread](https://www.reddit.com/r/StableDiffusion/comments/1h79ks2/fast_ltx_video_on_rtx_4060_and_other_ada_gpus/)
- **Diffusers integration:** A diffusers integration for the 8-bit model is already out! [Details here](https://github.com/sayakpaul/q8-ltx-video)
### TeaCache for LTX-Video 🍵
**TeaCache** is a training-free caching approach that leverages timestep differences across model outputs to accelerate LTX-Video inference by up to 2x without significant visual quality degradation.
- **Repository:** [TeaCache4LTX-Video](https://github.com/ali-vilab/TeaCache/tree/main/TeaCache4LTX-Video)
- **Features:**
- 🚀 Speeds up LTX-Video inference.
- 📊 Adjustable trade-offs between speed (up to 2x) and visual quality using configurable parameters.
- 🛠️ No retraining required: Works directly with existing models.
### Your Contribution
...is welcome! If you have a project or tool that integrates with LTX-Video,
please let us know by opening an issue or pull request.
# Training
We provide an open-source repository for fine-tuning the LTX-Video model: [LTX-Video-Trainer](https://github.com/Lightricks/LTX-Video-Trainer).
This repository supports both the 2B and 13B model variants, enabling full fine-tuning as well as LoRA (Low-Rank Adaptation) fine-tuning for more efficient training. This includes:
- **Control LoRAs**: Train custom control models like depth, pose, and canny control
- **Effect LoRAs**: Create specialized effects and transformations for video generation
Explore the repository to customize the model for your specific use cases!
More information and training instructions can be found in the [README](https://github.com/Lightricks/LTX-Video-Trainer/blob/main/README.md).
# Control Models
[ComfyUI-LTXVideo](https://github.com/Lightricks/ComfyUI-LTXVideo) repository now contains workflows and models for 3 specialized models that enable precise control over LTX-Video generation:
Pose Control, Depth Control and Canny Control
**Example ComfyUI Workflow (for all control types):** [ic-lora.json](https://github.com/Lightricks/ComfyUI-LTXVideo/blob/master/example_workflows/ic_lora/ic-lora.json)
# Join Us
Want to work on cutting-edge AI research and make a real impact on millions of users worldwide?
At **Lightricks**, an AI-first company, we're revolutionizing how visual content is created.
If you are passionate about AI, computer vision, and video generation, we would love to hear from you!
Please visit our [careers page](https://careers.lightricks.com/careers?query=&office=all&department=R%26D) for more information.
# Acknowledgement
We are grateful for the following awesome projects when implementing LTX-Video:
* [DiT](https://github.com/facebookresearch/DiT) and [PixArt-alpha](https://github.com/PixArt-alpha/PixArt-alpha): vision transformers for image generation.
## Citation
📄 Our tech report is out! If you find our work helpful, please ⭐️ star the repository and cite our paper.
```
@article{HaCohen2024LTXVideo,
title={LTX-Video: Realtime Video Latent Diffusion},
author={HaCohen, Yoav and Chiprut, Nisan and Brazowski, Benny and Shalem, Daniel and Moshe, Dudu and Richardson, Eitan and Levin, Eran and Shiran, Guy and Zabari, Nir and Gordon, Ori and Panet, Poriya and Weissbuch, Sapir and Kulikov, Victor and Bitterman, Yaki and Melumian, Zeev and Bibi, Ofir},
journal={arXiv preprint arXiv:2501.00103},
year={2024}
}
```
================================================
FILE: configs/ltxv-13b-0.9.8-dev-fp8.yaml
================================================
pipeline_type: multi-scale
checkpoint_path: "ltxv-13b-0.9.8-dev-fp8.safetensors"
downscale_factor: 0.6666666
spatial_upscaler_model_path: "ltxv-spatial-upscaler-0.9.8.safetensors"
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "float8_e4m3fn" # options: "float8_e4m3fn", "bfloat16", "mixed_precision"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
first_pass:
guidance_scale: [1, 1, 6, 8, 6, 1, 1]
stg_scale: [0, 0, 4, 4, 4, 2, 1]
rescaling_scale: [1, 1, 0.5, 0.5, 1, 1, 1]
guidance_timesteps: [1.0, 0.996, 0.9933, 0.9850, 0.9767, 0.9008, 0.6180]
skip_block_list: [[], [11, 25, 35, 39], [22, 35, 39], [28], [28], [28], [28]]
num_inference_steps: 30
skip_final_inference_steps: 3
cfg_star_rescale: true
second_pass:
guidance_scale: [1]
stg_scale: [1]
rescaling_scale: [1]
guidance_timesteps: [1.0]
skip_block_list: [27]
num_inference_steps: 30
skip_initial_inference_steps: 17
cfg_star_rescale: true
================================================
FILE: configs/ltxv-13b-0.9.8-dev.yaml
================================================
pipeline_type: multi-scale
checkpoint_path: "ltxv-13b-0.9.8-dev.safetensors"
downscale_factor: 0.6666666
spatial_upscaler_model_path: "ltxv-spatial-upscaler-0.9.8.safetensors"
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "bfloat16"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
first_pass:
guidance_scale: [1, 1, 6, 8, 6, 1, 1]
stg_scale: [0, 0, 4, 4, 4, 2, 1]
rescaling_scale: [1, 1, 0.5, 0.5, 1, 1, 1]
guidance_timesteps: [1.0, 0.996, 0.9933, 0.9850, 0.9767, 0.9008, 0.6180]
skip_block_list: [[], [11, 25, 35, 39], [22, 35, 39], [28], [28], [28], [28]]
num_inference_steps: 30
skip_final_inference_steps: 3
cfg_star_rescale: true
second_pass:
guidance_scale: [1]
stg_scale: [1]
rescaling_scale: [1]
guidance_timesteps: [1.0]
skip_block_list: [27]
num_inference_steps: 30
skip_initial_inference_steps: 17
cfg_star_rescale: true
================================================
FILE: configs/ltxv-13b-0.9.8-distilled-fp8.yaml
================================================
pipeline_type: multi-scale
checkpoint_path: "ltxv-13b-0.9.8-distilled-fp8.safetensors"
downscale_factor: 0.6666666
spatial_upscaler_model_path: "ltxv-spatial-upscaler-0.9.8.safetensors"
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "float8_e4m3fn" # options: "float8_e4m3fn", "bfloat16", "mixed_precision"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
first_pass:
timesteps: [1.0000, 0.9937, 0.9875, 0.9812, 0.9750, 0.9094, 0.7250]
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
skip_block_list: [42]
second_pass:
timesteps: [0.9094, 0.7250, 0.4219]
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
skip_block_list: [42]
tone_map_compression_ratio: 0.6
================================================
FILE: configs/ltxv-13b-0.9.8-distilled.yaml
================================================
pipeline_type: multi-scale
checkpoint_path: "ltxv-13b-0.9.8-distilled.safetensors"
downscale_factor: 0.6666666
spatial_upscaler_model_path: "ltxv-spatial-upscaler-0.9.8.safetensors"
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "bfloat16"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
first_pass:
timesteps: [1.0000, 0.9937, 0.9875, 0.9812, 0.9750, 0.9094, 0.7250]
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
skip_block_list: [42]
second_pass:
timesteps: [0.9094, 0.7250, 0.4219]
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
skip_block_list: [42]
tone_map_compression_ratio: 0.6
================================================
FILE: configs/ltxv-2b-0.9.1.yaml
================================================
pipeline_type: base
checkpoint_path: "ltx-video-2b-v0.9.1.safetensors"
guidance_scale: 3
stg_scale: 1
rescaling_scale: 0.7
skip_block_list: [19]
num_inference_steps: 40
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "bfloat16"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
================================================
FILE: configs/ltxv-2b-0.9.5.yaml
================================================
pipeline_type: base
checkpoint_path: "ltx-video-2b-v0.9.5.safetensors"
guidance_scale: 3
stg_scale: 1
rescaling_scale: 0.7
skip_block_list: [19]
num_inference_steps: 40
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "bfloat16"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
================================================
FILE: configs/ltxv-2b-0.9.6-dev.yaml
================================================
pipeline_type: base
checkpoint_path: "ltxv-2b-0.9.6-dev-04-25.safetensors"
guidance_scale: 3
stg_scale: 1
rescaling_scale: 0.7
skip_block_list: [19]
num_inference_steps: 40
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "bfloat16"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
================================================
FILE: configs/ltxv-2b-0.9.6-distilled.yaml
================================================
pipeline_type: base
checkpoint_path: "ltxv-2b-0.9.6-distilled-04-25.safetensors"
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
num_inference_steps: 8
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "bfloat16"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: true
================================================
FILE: configs/ltxv-2b-0.9.8-distilled-fp8.yaml
================================================
pipeline_type: multi-scale
checkpoint_path: "ltxv-2b-0.9.8-distilled-fp8.safetensors"
downscale_factor: 0.6666666
spatial_upscaler_model_path: "ltxv-spatial-upscaler-0.9.8.safetensors"
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "float8_e4m3fn" # options: "float8_e4m3fn", "bfloat16", "mixed_precision"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
first_pass:
timesteps: [1.0000, 0.9937, 0.9875, 0.9812, 0.9750, 0.9094, 0.7250]
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
skip_block_list: [42]
second_pass:
timesteps: [0.9094, 0.7250, 0.4219]
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
skip_block_list: [42]
================================================
FILE: configs/ltxv-2b-0.9.8-distilled.yaml
================================================
pipeline_type: multi-scale
checkpoint_path: "ltxv-2b-0.9.8-distilled.safetensors"
downscale_factor: 0.6666666
spatial_upscaler_model_path: "ltxv-spatial-upscaler-0.9.8.safetensors"
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "bfloat16"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
first_pass:
timesteps: [1.0000, 0.9937, 0.9875, 0.9812, 0.9750, 0.9094, 0.7250]
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
skip_block_list: [42]
second_pass:
timesteps: [0.9094, 0.7250, 0.4219]
guidance_scale: 1
stg_scale: 0
rescaling_scale: 1
skip_block_list: [42]
================================================
FILE: configs/ltxv-2b-0.9.yaml
================================================
pipeline_type: base
checkpoint_path: "ltx-video-2b-v0.9.safetensors"
guidance_scale: 3
stg_scale: 1
rescaling_scale: 0.7
skip_block_list: [19]
num_inference_steps: 40
stg_mode: "attention_values" # options: "attention_values", "attention_skip", "residual", "transformer_block"
decode_timestep: 0.05
decode_noise_scale: 0.025
text_encoder_model_name_or_path: "PixArt-alpha/PixArt-XL-2-1024-MS"
precision: "bfloat16"
sampler: "from_checkpoint" # options: "uniform", "linear-quadratic", "from_checkpoint"
prompt_enhancement_words_threshold: 120
prompt_enhancer_image_caption_model_name_or_path: "MiaoshouAI/Florence-2-large-PromptGen-v2.0"
prompt_enhancer_llm_model_name_or_path: "unsloth/Llama-3.2-3B-Instruct"
stochastic_sampling: false
================================================
FILE: inference.py
================================================
from transformers import HfArgumentParser
from ltx_video.inference import infer, InferenceConfig
def main():
parser = HfArgumentParser(InferenceConfig)
config = parser.parse_args_into_dataclasses()[0]
infer(config=config)
if __name__ == "__main__":
main()
================================================
FILE: ltx_video/__init__.py
================================================
================================================
FILE: ltx_video/inference.py
================================================
import os
import random
from datetime import datetime
from pathlib import Path
from diffusers.utils import logging
from typing import Optional, List, Union
import yaml
import imageio
import json
import numpy as np
import torch
from safetensors import safe_open
from PIL import Image
import torchvision.transforms.functional as TVF
from transformers import (
T5EncoderModel,
T5Tokenizer,
AutoModelForCausalLM,
AutoProcessor,
AutoTokenizer,
)
from huggingface_hub import hf_hub_download
from dataclasses import dataclass, field
from ltx_video.models.autoencoders.causal_video_autoencoder import (
CausalVideoAutoencoder,
)
from ltx_video.models.transformers.symmetric_patchifier import SymmetricPatchifier
from ltx_video.models.transformers.transformer3d import Transformer3DModel
from ltx_video.pipelines.pipeline_ltx_video import (
ConditioningItem,
LTXVideoPipeline,
LTXMultiScalePipeline,
)
from ltx_video.schedulers.rf import RectifiedFlowScheduler
from ltx_video.utils.skip_layer_strategy import SkipLayerStrategy
from ltx_video.models.autoencoders.latent_upsampler import LatentUpsampler
import ltx_video.pipelines.crf_compressor as crf_compressor
logger = logging.get_logger("LTX-Video")
def get_total_gpu_memory():
if torch.cuda.is_available():
total_memory = torch.cuda.get_device_properties(0).total_memory / (1024**3)
return total_memory
return 0
def get_device():
if torch.cuda.is_available():
return "cuda"
elif torch.backends.mps.is_available():
return "mps"
return "cpu"
def load_image_to_tensor_with_resize_and_crop(
image_input: Union[str, Image.Image],
target_height: int = 512,
target_width: int = 768,
just_crop: bool = False,
) -> torch.Tensor:
"""Load and process an image into a tensor.
Args:
image_input: Either a file path (str) or a PIL Image object
target_height: Desired height of output tensor
target_width: Desired width of output tensor
just_crop: If True, only crop the image to the target size without resizing
"""
if isinstance(image_input, str):
image = Image.open(image_input).convert("RGB")
elif isinstance(image_input, Image.Image):
image = image_input
else:
raise ValueError("image_input must be either a file path or a PIL Image object")
input_width, input_height = image.size
aspect_ratio_target = target_width / target_height
aspect_ratio_frame = input_width / input_height
if aspect_ratio_frame > aspect_ratio_target:
new_width = int(input_height * aspect_ratio_target)
new_height = input_height
x_start = (input_width - new_width) // 2
y_start = 0
else:
new_width = input_width
new_height = int(input_width / aspect_ratio_target)
x_start = 0
y_start = (input_height - new_height) // 2
image = image.crop((x_start, y_start, x_start + new_width, y_start + new_height))
if not just_crop:
image = image.resize((target_width, target_height))
frame_tensor = TVF.to_tensor(image) # PIL -> tensor (C, H, W), [0,1]
frame_tensor = TVF.gaussian_blur(frame_tensor, kernel_size=3, sigma=1.0)
frame_tensor_hwc = frame_tensor.permute(1, 2, 0) # (C, H, W) -> (H, W, C)
frame_tensor_hwc = crf_compressor.compress(frame_tensor_hwc)
frame_tensor = frame_tensor_hwc.permute(2, 0, 1) * 255.0 # (H, W, C) -> (C, H, W)
frame_tensor = (frame_tensor / 127.5) - 1.0
# Create 5D tensor: (batch_size=1, channels=3, num_frames=1, height, width)
return frame_tensor.unsqueeze(0).unsqueeze(2)
def calculate_padding(
source_height: int, source_width: int, target_height: int, target_width: int
) -> tuple[int, int, int, int]:
# Calculate total padding needed
pad_height = target_height - source_height
pad_width = target_width - source_width
# Calculate padding for each side
pad_top = pad_height // 2
pad_bottom = pad_height - pad_top # Handles odd padding
pad_left = pad_width // 2
pad_right = pad_width - pad_left # Handles odd padding
# Return padded tensor
# Padding format is (left, right, top, bottom)
padding = (pad_left, pad_right, pad_top, pad_bottom)
return padding
def convert_prompt_to_filename(text: str, max_len: int = 20) -> str:
# Remove non-letters and convert to lowercase
clean_text = "".join(
char.lower() for char in text if char.isalpha() or char.isspace()
)
# Split into words
words = clean_text.split()
# Build result string keeping track of length
result = []
current_length = 0
for word in words:
# Add word length plus 1 for underscore (except for first word)
new_length = current_length + len(word)
if new_length <= max_len:
result.append(word)
current_length += len(word)
else:
break
return "-".join(result)
# Generate output video name
def get_unique_filename(
base: str,
ext: str,
prompt: str,
seed: int,
resolution: tuple[int, int, int],
dir: Path,
endswith=None,
index_range=1000,
) -> Path:
base_filename = f"{base}_{convert_prompt_to_filename(prompt, max_len=30)}_{seed}_{resolution[0]}x{resolution[1]}x{resolution[2]}"
for i in range(index_range):
filename = dir / f"{base_filename}_{i}{endswith if endswith else ''}{ext}"
if not os.path.exists(filename):
return filename
raise FileExistsError(
f"Could not find a unique filename after {index_range} attempts."
)
def seed_everething(seed: int):
random.seed(seed)
np.random.seed(seed)
torch.manual_seed(seed)
if torch.cuda.is_available():
torch.cuda.manual_seed(seed)
if torch.backends.mps.is_available():
torch.mps.manual_seed(seed)
def create_transformer(ckpt_path: str, precision: str) -> Transformer3DModel:
if precision == "float8_e4m3fn":
try:
from q8_kernels.integration.patch_transformer import (
patch_diffusers_transformer as patch_transformer_for_q8_kernels,
)
transformer = Transformer3DModel.from_pretrained(
ckpt_path, dtype=torch.float8_e4m3fn
)
patch_transformer_for_q8_kernels(transformer)
return transformer
except ImportError:
raise ValueError(
"Q8-Kernels not found. To use FP8 checkpoint, please install Q8 kernels from https://github.com/Lightricks/LTXVideo-Q8-Kernels"
)
elif precision == "bfloat16":
return Transformer3DModel.from_pretrained(ckpt_path).to(torch.bfloat16)
else:
return Transformer3DModel.from_pretrained(ckpt_path)
def create_ltx_video_pipeline(
ckpt_path: str,
precision: str,
text_encoder_model_name_or_path: str,
sampler: Optional[str] = None,
device: Optional[str] = None,
enhance_prompt: bool = False,
prompt_enhancer_image_caption_model_name_or_path: Optional[str] = None,
prompt_enhancer_llm_model_name_or_path: Optional[str] = None,
) -> LTXVideoPipeline:
ckpt_path = Path(ckpt_path)
assert os.path.exists(
ckpt_path
), f"Ckpt path provided (--ckpt_path) {ckpt_path} does not exist"
with safe_open(ckpt_path, framework="pt") as f:
metadata = f.metadata()
config_str = metadata.get("config")
configs = json.loads(config_str)
allowed_inference_steps = configs.get("allowed_inference_steps", None)
vae = CausalVideoAutoencoder.from_pretrained(ckpt_path)
transformer = create_transformer(ckpt_path, precision)
# Use constructor if sampler is specified, otherwise use from_pretrained
if sampler == "from_checkpoint" or not sampler:
scheduler = RectifiedFlowScheduler.from_pretrained(ckpt_path)
else:
scheduler = RectifiedFlowScheduler(
sampler=("Uniform" if sampler.lower() == "uniform" else "LinearQuadratic")
)
text_encoder = T5EncoderModel.from_pretrained(
text_encoder_model_name_or_path, subfolder="text_encoder"
)
patchifier = SymmetricPatchifier(patch_size=1)
tokenizer = T5Tokenizer.from_pretrained(
text_encoder_model_name_or_path, subfolder="tokenizer"
)
transformer = transformer.to(device)
vae = vae.to(device)
text_encoder = text_encoder.to(device)
if enhance_prompt:
prompt_enhancer_image_caption_model = AutoModelForCausalLM.from_pretrained(
prompt_enhancer_image_caption_model_name_or_path, trust_remote_code=True
)
prompt_enhancer_image_caption_processor = AutoProcessor.from_pretrained(
prompt_enhancer_image_caption_model_name_or_path, trust_remote_code=True
)
prompt_enhancer_llm_model = AutoModelForCausalLM.from_pretrained(
prompt_enhancer_llm_model_name_or_path,
torch_dtype="bfloat16",
)
prompt_enhancer_llm_tokenizer = AutoTokenizer.from_pretrained(
prompt_enhancer_llm_model_name_or_path,
)
else:
prompt_enhancer_image_caption_model = None
prompt_enhancer_image_caption_processor = None
prompt_enhancer_llm_model = None
prompt_enhancer_llm_tokenizer = None
vae = vae.to(torch.bfloat16)
text_encoder = text_encoder.to(torch.bfloat16)
# Use submodels for the pipeline
submodel_dict = {
"transformer": transformer,
"patchifier": patchifier,
"text_encoder": text_encoder,
"tokenizer": tokenizer,
"scheduler": scheduler,
"vae": vae,
"prompt_enhancer_image_caption_model": prompt_enhancer_image_caption_model,
"prompt_enhancer_image_caption_processor": prompt_enhancer_image_caption_processor,
"prompt_enhancer_llm_model": prompt_enhancer_llm_model,
"prompt_enhancer_llm_tokenizer": prompt_enhancer_llm_tokenizer,
"allowed_inference_steps": allowed_inference_steps,
}
pipeline = LTXVideoPipeline(**submodel_dict)
pipeline = pipeline.to(device)
return pipeline
def create_latent_upsampler(latent_upsampler_model_path: str, device: str):
latent_upsampler = LatentUpsampler.from_pretrained(latent_upsampler_model_path)
latent_upsampler.to(device)
latent_upsampler.eval()
return latent_upsampler
def load_pipeline_config(pipeline_config: str):
current_file = Path(__file__)
path = None
if os.path.isfile(current_file.parent / pipeline_config):
path = current_file.parent / pipeline_config
elif os.path.isfile(pipeline_config):
path = pipeline_config
else:
raise ValueError(f"Pipeline config file {pipeline_config} does not exist")
with open(path, "r") as f:
return yaml.safe_load(f)
@dataclass
class InferenceConfig:
prompt: str = field(metadata={"help": "Prompt for the generation"})
output_path: str = field(
default_factory=lambda: Path(
f"outputs/{datetime.today().strftime('%Y-%m-%d')}"
),
metadata={"help": "Path to the folder to save the output video"},
)
# Pipeline settings
pipeline_config: str = field(
default="configs/ltxv-13b-0.9.7-dev.yaml",
metadata={"help": "Path to the pipeline config file"},
)
seed: int = field(
default=171198, metadata={"help": "Random seed for the inference"}
)
height: int = field(
default=704, metadata={"help": "Height of the output video frames"}
)
width: int = field(
default=1216, metadata={"help": "Width of the output video frames"}
)
num_frames: int = field(
default=121,
metadata={"help": "Number of frames to generate in the output video"},
)
frame_rate: int = field(
default=30, metadata={"help": "Frame rate for the output video"}
)
offload_to_cpu: bool = field(
default=False, metadata={"help": "Offloading unnecessary computations to CPU."}
)
negative_prompt: str = field(
default="worst quality, inconsistent motion, blurry, jittery, distorted",
metadata={"help": "Negative prompt for undesired features"},
)
# Video-to-video arguments
input_media_path: Optional[str] = field(
default=None,
metadata={
"help": "Path to the input video (or image) to be modified using the video-to-video pipeline"
},
)
# Conditioning
image_cond_noise_scale: float = field(
default=0.15,
metadata={"help": "Amount of noise to add to the conditioned image"},
)
conditioning_media_paths: Optional[List[str]] = field(
default=None,
metadata={
"help": "List of paths to conditioning media (images or videos). Each path will be used as a conditioning item."
},
)
conditioning_strengths: Optional[List[float]] = field(
default=None,
metadata={
"help": "List of conditioning strengths (between 0 and 1) for each conditioning item. Must match the number of conditioning items."
},
)
conditioning_start_frames: Optional[List[int]] = field(
default=None,
metadata={
"help": "List of frame indices where each conditioning item should be applied. Must match the number of conditioning items."
},
)
def infer(config: InferenceConfig):
pipeline_config = load_pipeline_config(config.pipeline_config)
ltxv_model_name_or_path = pipeline_config["checkpoint_path"]
if not os.path.isfile(ltxv_model_name_or_path):
ltxv_model_path = hf_hub_download(
repo_id="Lightricks/LTX-Video",
filename=ltxv_model_name_or_path,
repo_type="model",
)
else:
ltxv_model_path = ltxv_model_name_or_path
spatial_upscaler_model_name_or_path = pipeline_config.get(
"spatial_upscaler_model_path"
)
if spatial_upscaler_model_name_or_path and not os.path.isfile(
spatial_upscaler_model_name_or_path
):
spatial_upscaler_model_path = hf_hub_download(
repo_id="Lightricks/LTX-Video",
filename=spatial_upscaler_model_name_or_path,
repo_type="model",
)
else:
spatial_upscaler_model_path = spatial_upscaler_model_name_or_path
conditioning_media_paths = config.conditioning_media_paths
conditioning_strengths = config.conditioning_strengths
conditioning_start_frames = config.conditioning_start_frames
# Validate conditioning arguments
if conditioning_media_paths:
# Use default strengths of 1.0
if not conditioning_strengths:
conditioning_strengths = [1.0] * len(conditioning_media_paths)
if not conditioning_start_frames:
raise ValueError(
"If `conditioning_media_paths` is provided, "
"`conditioning_start_frames` must also be provided"
)
if len(conditioning_media_paths) != len(conditioning_strengths) or len(
conditioning_media_paths
) != len(conditioning_start_frames):
raise ValueError(
"`conditioning_media_paths`, `conditioning_strengths`, "
"and `conditioning_start_frames` must have the same length"
)
if any(s < 0 or s > 1 for s in conditioning_strengths):
raise ValueError("All conditioning strengths must be between 0 and 1")
if any(f < 0 or f >= config.num_frames for f in conditioning_start_frames):
raise ValueError(
f"All conditioning start frames must be between 0 and {config.num_frames-1}"
)
seed_everething(config.seed)
if config.offload_to_cpu and not torch.cuda.is_available():
logger.warning(
"offload_to_cpu is set to True, but offloading will not occur since the model is already running on CPU."
)
offload_to_cpu = False
else:
offload_to_cpu = config.offload_to_cpu and get_total_gpu_memory() < 30
output_dir = (
Path(config.output_path)
if config.output_path
else Path(f"outputs/{datetime.today().strftime('%Y-%m-%d')}")
)
output_dir.mkdir(parents=True, exist_ok=True)
# Adjust dimensions to be divisible by 32 and num_frames to be (N * 8 + 1)
height_padded = ((config.height - 1) // 32 + 1) * 32
width_padded = ((config.width - 1) // 32 + 1) * 32
num_frames_padded = ((config.num_frames - 2) // 8 + 1) * 8 + 1
padding = calculate_padding(
config.height, config.width, height_padded, width_padded
)
logger.warning(
f"Padded dimensions: {height_padded}x{width_padded}x{num_frames_padded}"
)
device = get_device()
prompt_enhancement_words_threshold = pipeline_config[
"prompt_enhancement_words_threshold"
]
prompt_word_count = len(config.prompt.split())
enhance_prompt = (
prompt_enhancement_words_threshold > 0
and prompt_word_count < prompt_enhancement_words_threshold
)
if prompt_enhancement_words_threshold > 0 and not enhance_prompt:
logger.info(
f"Prompt has {prompt_word_count} words, which exceeds the threshold of {prompt_enhancement_words_threshold}. Prompt enhancement disabled."
)
precision = pipeline_config["precision"]
text_encoder_model_name_or_path = pipeline_config["text_encoder_model_name_or_path"]
sampler = pipeline_config.get("sampler", None)
prompt_enhancer_image_caption_model_name_or_path = pipeline_config[
"prompt_enhancer_image_caption_model_name_or_path"
]
prompt_enhancer_llm_model_name_or_path = pipeline_config[
"prompt_enhancer_llm_model_name_or_path"
]
pipeline = create_ltx_video_pipeline(
ckpt_path=ltxv_model_path,
precision=precision,
text_encoder_model_name_or_path=text_encoder_model_name_or_path,
sampler=sampler,
device=device,
enhance_prompt=enhance_prompt,
prompt_enhancer_image_caption_model_name_or_path=prompt_enhancer_image_caption_model_name_or_path,
prompt_enhancer_llm_model_name_or_path=prompt_enhancer_llm_model_name_or_path,
)
if pipeline_config.get("pipeline_type", None) == "multi-scale":
if not spatial_upscaler_model_path:
raise ValueError(
"spatial upscaler model path is missing from pipeline config file and is required for multi-scale rendering"
)
latent_upsampler = create_latent_upsampler(
spatial_upscaler_model_path, pipeline.device
)
pipeline = LTXMultiScalePipeline(pipeline, latent_upsampler=latent_upsampler)
media_item = None
if config.input_media_path:
media_item = load_media_file(
media_path=config.input_media_path,
height=config.height,
width=config.width,
max_frames=num_frames_padded,
padding=padding,
)
conditioning_items = (
prepare_conditioning(
conditioning_media_paths=conditioning_media_paths,
conditioning_strengths=conditioning_strengths,
conditioning_start_frames=conditioning_start_frames,
height=config.height,
width=config.width,
num_frames=config.num_frames,
padding=padding,
pipeline=pipeline,
)
if conditioning_media_paths
else None
)
stg_mode = pipeline_config.get("stg_mode", "attention_values")
del pipeline_config["stg_mode"]
if stg_mode.lower() == "stg_av" or stg_mode.lower() == "attention_values":
skip_layer_strategy = SkipLayerStrategy.AttentionValues
elif stg_mode.lower() == "stg_as" or stg_mode.lower() == "attention_skip":
skip_layer_strategy = SkipLayerStrategy.AttentionSkip
elif stg_mode.lower() == "stg_r" or stg_mode.lower() == "residual":
skip_layer_strategy = SkipLayerStrategy.Residual
elif stg_mode.lower() == "stg_t" or stg_mode.lower() == "transformer_block":
skip_layer_strategy = SkipLayerStrategy.TransformerBlock
else:
raise ValueError(f"Invalid spatiotemporal guidance mode: {stg_mode}")
# Prepare input for the pipeline
sample = {
"prompt": config.prompt,
"prompt_attention_mask": None,
"negative_prompt": config.negative_prompt,
"negative_prompt_attention_mask": None,
}
generator = torch.Generator(device=device).manual_seed(config.seed)
images = pipeline(
**pipeline_config,
skip_layer_strategy=skip_layer_strategy,
generator=generator,
output_type="pt",
callback_on_step_end=None,
height=height_padded,
width=width_padded,
num_frames=num_frames_padded,
frame_rate=config.frame_rate,
**sample,
media_items=media_item,
conditioning_items=conditioning_items,
is_video=True,
vae_per_channel_normalize=True,
image_cond_noise_scale=config.image_cond_noise_scale,
mixed_precision=(precision == "mixed_precision"),
offload_to_cpu=offload_to_cpu,
device=device,
enhance_prompt=enhance_prompt,
).images
# Crop the padded images to the desired resolution and number of frames
(pad_left, pad_right, pad_top, pad_bottom) = padding
pad_bottom = -pad_bottom
pad_right = -pad_right
if pad_bottom == 0:
pad_bottom = images.shape[3]
if pad_right == 0:
pad_right = images.shape[4]
images = images[:, :, : config.num_frames, pad_top:pad_bottom, pad_left:pad_right]
for i in range(images.shape[0]):
# Gathering from B, C, F, H, W to C, F, H, W and then permuting to F, H, W, C
video_np = images[i].permute(1, 2, 3, 0).cpu().float().numpy()
# Unnormalizing images to [0, 255] range
video_np = (video_np * 255).astype(np.uint8)
fps = config.frame_rate
height, width = video_np.shape[1:3]
# In case a single image is generated
if video_np.shape[0] == 1:
output_filename = get_unique_filename(
f"image_output_{i}",
".png",
prompt=config.prompt,
seed=config.seed,
resolution=(height, width, config.num_frames),
dir=output_dir,
)
imageio.imwrite(output_filename, video_np[0])
else:
output_filename = get_unique_filename(
f"video_output_{i}",
".mp4",
prompt=config.prompt,
seed=config.seed,
resolution=(height, width, config.num_frames),
dir=output_dir,
)
# Write video
with imageio.get_writer(output_filename, fps=fps) as video:
for frame in video_np:
video.append_data(frame)
logger.warning(f"Output saved to {output_filename}")
def prepare_conditioning(
conditioning_media_paths: List[str],
conditioning_strengths: List[float],
conditioning_start_frames: List[int],
height: int,
width: int,
num_frames: int,
padding: tuple[int, int, int, int],
pipeline: LTXVideoPipeline,
) -> Optional[List[ConditioningItem]]:
"""Prepare conditioning items based on input media paths and their parameters.
Args:
conditioning_media_paths: List of paths to conditioning media (images or videos)
conditioning_strengths: List of conditioning strengths for each media item
conditioning_start_frames: List of frame indices where each item should be applied
height: Height of the output frames
width: Width of the output frames
num_frames: Number of frames in the output video
padding: Padding to apply to the frames
pipeline: LTXVideoPipeline object used for condition video trimming
Returns:
A list of ConditioningItem objects.
"""
conditioning_items = []
for path, strength, start_frame in zip(
conditioning_media_paths, conditioning_strengths, conditioning_start_frames
):
num_input_frames = orig_num_input_frames = get_media_num_frames(path)
if hasattr(pipeline, "trim_conditioning_sequence") and callable(
getattr(pipeline, "trim_conditioning_sequence")
):
num_input_frames = pipeline.trim_conditioning_sequence(
start_frame, orig_num_input_frames, num_frames
)
if num_input_frames < orig_num_input_frames:
logger.warning(
f"Trimming conditioning video {path} from {orig_num_input_frames} to {num_input_frames} frames."
)
media_tensor = load_media_file(
media_path=path,
height=height,
width=width,
max_frames=num_input_frames,
padding=padding,
just_crop=True,
)
conditioning_items.append(ConditioningItem(media_tensor, start_frame, strength))
return conditioning_items
def get_media_num_frames(media_path: str) -> int:
is_video = any(
media_path.lower().endswith(ext) for ext in [".mp4", ".avi", ".mov", ".mkv"]
)
num_frames = 1
if is_video:
reader = imageio.get_reader(media_path)
num_frames = reader.count_frames()
reader.close()
return num_frames
def load_media_file(
media_path: str,
height: int,
width: int,
max_frames: int,
padding: tuple[int, int, int, int],
just_crop: bool = False,
) -> torch.Tensor:
is_video = any(
media_path.lower().endswith(ext) for ext in [".mp4", ".avi", ".mov", ".mkv"]
)
if is_video:
reader = imageio.get_reader(media_path)
num_input_frames = min(reader.count_frames(), max_frames)
# Read and preprocess the relevant frames from the video file.
frames = []
for i in range(num_input_frames):
frame = Image.fromarray(reader.get_data(i))
frame_tensor = load_image_to_tensor_with_resize_and_crop(
frame, height, width, just_crop=just_crop
)
frame_tensor = torch.nn.functional.pad(frame_tensor, padding)
frames.append(frame_tensor)
reader.close()
# Stack frames along the temporal dimension
media_tensor = torch.cat(frames, dim=2)
else: # Input image
media_tensor = load_image_to_tensor_with_resize_and_crop(
media_path, height, width, just_crop=just_crop
)
media_tensor = torch.nn.functional.pad(media_tensor, padding)
return media_tensor
================================================
FILE: ltx_video/models/__init__.py
================================================
================================================
FILE: ltx_video/models/autoencoders/__init__.py
================================================
================================================
FILE: ltx_video/models/autoencoders/causal_conv3d.py
================================================
from typing import Tuple, Union
import torch
import torch.nn as nn
class CausalConv3d(nn.Module):
def __init__(
self,
in_channels,
out_channels,
kernel_size: int = 3,
stride: Union[int, Tuple[int]] = 1,
dilation: int = 1,
groups: int = 1,
spatial_padding_mode: str = "zeros",
**kwargs,
):
super().__init__()
self.in_channels = in_channels
self.out_channels = out_channels
kernel_size = (kernel_size, kernel_size, kernel_size)
self.time_kernel_size = kernel_size[0]
dilation = (dilation, 1, 1)
height_pad = kernel_size[1] // 2
width_pad = kernel_size[2] // 2
padding = (0, height_pad, width_pad)
self.conv = nn.Conv3d(
in_channels,
out_channels,
kernel_size,
stride=stride,
dilation=dilation,
padding=padding,
padding_mode=spatial_padding_mode,
groups=groups,
)
def forward(self, x, causal: bool = True):
if causal:
first_frame_pad = x[:, :, :1, :, :].repeat(
(1, 1, self.time_kernel_size - 1, 1, 1)
)
x = torch.concatenate((first_frame_pad, x), dim=2)
else:
first_frame_pad = x[:, :, :1, :, :].repeat(
(1, 1, (self.time_kernel_size - 1) // 2, 1, 1)
)
last_frame_pad = x[:, :, -1:, :, :].repeat(
(1, 1, (self.time_kernel_size - 1) // 2, 1, 1)
)
x = torch.concatenate((first_frame_pad, x, last_frame_pad), dim=2)
x = self.conv(x)
return x
@property
def weight(self):
return self.conv.weight
================================================
FILE: ltx_video/models/autoencoders/causal_video_autoencoder.py
================================================
import json
import os
from functools import partial
from types import SimpleNamespace
from typing import Any, Mapping, Optional, Tuple, Union, List
from pathlib import Path
import torch
import numpy as np
from einops import rearrange
from torch import nn
from diffusers.utils import logging
import torch.nn.functional as F
from diffusers.models.embeddings import PixArtAlphaCombinedTimestepSizeEmbeddings
from safetensors import safe_open
from ltx_video.models.autoencoders.conv_nd_factory import make_conv_nd, make_linear_nd
from ltx_video.models.autoencoders.pixel_norm import PixelNorm
from ltx_video.models.autoencoders.pixel_shuffle import PixelShuffleND
from ltx_video.models.autoencoders.vae import AutoencoderKLWrapper
from ltx_video.models.transformers.attention import Attention
from ltx_video.utils.diffusers_config_mapping import (
diffusers_and_ours_config_mapping,
make_hashable_key,
VAE_KEYS_RENAME_DICT,
)
PER_CHANNEL_STATISTICS_PREFIX = "per_channel_statistics."
logger = logging.get_logger(__name__) # pylint: disable=invalid-name
class CausalVideoAutoencoder(AutoencoderKLWrapper):
@classmethod
def from_pretrained(
cls,
pretrained_model_name_or_path: Optional[Union[str, os.PathLike]],
*args,
**kwargs,
):
pretrained_model_name_or_path = Path(pretrained_model_name_or_path)
if (
pretrained_model_name_or_path.is_dir()
and (pretrained_model_name_or_path / "autoencoder.pth").exists()
):
config_local_path = pretrained_model_name_or_path / "config.json"
config = cls.load_config(config_local_path, **kwargs)
model_local_path = pretrained_model_name_or_path / "autoencoder.pth"
state_dict = torch.load(model_local_path, map_location=torch.device("cpu"))
statistics_local_path = (
pretrained_model_name_or_path / "per_channel_statistics.json"
)
if statistics_local_path.exists():
with open(statistics_local_path, "r") as file:
data = json.load(file)
transposed_data = list(zip(*data["data"]))
data_dict = {
col: torch.tensor(vals)
for col, vals in zip(data["columns"], transposed_data)
}
std_of_means = data_dict["std-of-means"]
mean_of_means = data_dict.get(
"mean-of-means", torch.zeros_like(data_dict["std-of-means"])
)
state_dict[f"{PER_CHANNEL_STATISTICS_PREFIX}std-of-means"] = (
std_of_means
)
state_dict[f"{PER_CHANNEL_STATISTICS_PREFIX}mean-of-means"] = (
mean_of_means
)
elif pretrained_model_name_or_path.is_dir():
config_path = pretrained_model_name_or_path / "vae" / "config.json"
with open(config_path, "r") as f:
config = make_hashable_key(json.load(f))
assert config in diffusers_and_ours_config_mapping, (
"Provided diffusers checkpoint config for VAE is not suppported. "
"We only support diffusers configs found in Lightricks/LTX-Video."
)
config = diffusers_and_ours_config_mapping[config]
state_dict_path = (
pretrained_model_name_or_path
/ "vae"
/ "diffusion_pytorch_model.safetensors"
)
state_dict = {}
with safe_open(state_dict_path, framework="pt", device="cpu") as f:
for k in f.keys():
state_dict[k] = f.get_tensor(k)
for key in list(state_dict.keys()):
new_key = key
for replace_key, rename_key in VAE_KEYS_RENAME_DICT.items():
new_key = new_key.replace(replace_key, rename_key)
state_dict[new_key] = state_dict.pop(key)
elif pretrained_model_name_or_path.is_file() and str(
pretrained_model_name_or_path
).endswith(".safetensors"):
state_dict = {}
with safe_open(
pretrained_model_name_or_path, framework="pt", device="cpu"
) as f:
metadata = f.metadata()
for k in f.keys():
state_dict[k] = f.get_tensor(k)
configs = json.loads(metadata["config"])
config = configs["vae"]
video_vae = cls.from_config(config)
if "torch_dtype" in kwargs:
video_vae.to(kwargs["torch_dtype"])
video_vae.load_state_dict(state_dict)
return video_vae
@staticmethod
def from_config(config):
assert (
config["_class_name"] == "CausalVideoAutoencoder"
), "config must have _class_name=CausalVideoAutoencoder"
if isinstance(config["dims"], list):
config["dims"] = tuple(config["dims"])
assert config["dims"] in [2, 3, (2, 1)], "dims must be 2, 3 or (2, 1)"
double_z = config.get("double_z", True)
latent_log_var = config.get(
"latent_log_var", "per_channel" if double_z else "none"
)
use_quant_conv = config.get("use_quant_conv", True)
normalize_latent_channels = config.get("normalize_latent_channels", False)
if use_quant_conv and latent_log_var in ["uniform", "constant"]:
raise ValueError(
f"latent_log_var={latent_log_var} requires use_quant_conv=False"
)
encoder = Encoder(
dims=config["dims"],
in_channels=config.get("in_channels", 3),
out_channels=config["latent_channels"],
blocks=config.get("encoder_blocks", config.get("blocks")),
patch_size=config.get("patch_size", 1),
latent_log_var=latent_log_var,
norm_layer=config.get("norm_layer", "group_norm"),
base_channels=config.get("encoder_base_channels", 128),
spatial_padding_mode=config.get("spatial_padding_mode", "zeros"),
)
decoder = Decoder(
dims=config["dims"],
in_channels=config["latent_channels"],
out_channels=config.get("out_channels", 3),
blocks=config.get("decoder_blocks", config.get("blocks")),
patch_size=config.get("patch_size", 1),
norm_layer=config.get("norm_layer", "group_norm"),
causal=config.get("causal_decoder", False),
timestep_conditioning=config.get("timestep_conditioning", False),
base_channels=config.get("decoder_base_channels", 128),
spatial_padding_mode=config.get("spatial_padding_mode", "zeros"),
)
dims = config["dims"]
return CausalVideoAutoencoder(
encoder=encoder,
decoder=decoder,
latent_channels=config["latent_channels"],
dims=dims,
use_quant_conv=use_quant_conv,
normalize_latent_channels=normalize_latent_channels,
)
@property
def config(self):
return SimpleNamespace(
_class_name="CausalVideoAutoencoder",
dims=self.dims,
in_channels=self.encoder.conv_in.in_channels // self.encoder.patch_size**2,
out_channels=self.decoder.conv_out.out_channels
// self.decoder.patch_size**2,
latent_channels=self.decoder.conv_in.in_channels,
encoder_blocks=self.encoder.blocks_desc,
decoder_blocks=self.decoder.blocks_desc,
scaling_factor=1.0,
norm_layer=self.encoder.norm_layer,
patch_size=self.encoder.patch_size,
latent_log_var=self.encoder.latent_log_var,
use_quant_conv=self.use_quant_conv,
causal_decoder=self.decoder.causal,
timestep_conditioning=self.decoder.timestep_conditioning,
normalize_latent_channels=self.normalize_latent_channels,
)
@property
def is_video_supported(self):
"""
Check if the model supports video inputs of shape (B, C, F, H, W). Otherwise, the model only supports 2D images.
"""
return self.dims != 2
@property
def spatial_downscale_factor(self):
return (
2
** len(
[
block
for block in self.encoder.blocks_desc
if block[0]
in [
"compress_space",
"compress_all",
"compress_all_res",
"compress_space_res",
]
]
)
* self.encoder.patch_size
)
@property
def temporal_downscale_factor(self):
return 2 ** len(
[
block
for block in self.encoder.blocks_desc
if block[0]
in [
"compress_time",
"compress_all",
"compress_all_res",
"compress_time_res",
]
]
)
def to_json_string(self) -> str:
import json
return json.dumps(self.config.__dict__)
def load_state_dict(self, state_dict: Mapping[str, Any], strict: bool = True):
if any([key.startswith("vae.") for key in state_dict.keys()]):
state_dict = {
key.replace("vae.", ""): value
for key, value in state_dict.items()
if key.startswith("vae.")
}
ckpt_state_dict = {
key: value
for key, value in state_dict.items()
if not key.startswith(PER_CHANNEL_STATISTICS_PREFIX)
}
model_keys = set(name for name, _ in self.named_modules())
key_mapping = {
".resnets.": ".res_blocks.",
"downsamplers.0": "downsample",
"upsamplers.0": "upsample",
}
converted_state_dict = {}
for key, value in ckpt_state_dict.items():
for k, v in key_mapping.items():
key = key.replace(k, v)
key_prefix = ".".join(key.split(".")[:-1])
if "norm" in key and key_prefix not in model_keys:
logger.info(
f"Removing key {key} from state_dict as it is not present in the model"
)
continue
converted_state_dict[key] = value
super().load_state_dict(converted_state_dict, strict=strict)
data_dict = {
key.removeprefix(PER_CHANNEL_STATISTICS_PREFIX): value
for key, value in state_dict.items()
if key.startswith(PER_CHANNEL_STATISTICS_PREFIX)
}
if len(data_dict) > 0:
self.register_buffer("std_of_means", data_dict["std-of-means"])
self.register_buffer(
"mean_of_means",
data_dict.get(
"mean-of-means", torch.zeros_like(data_dict["std-of-means"])
),
)
def last_layer(self):
if hasattr(self.decoder, "conv_out"):
if isinstance(self.decoder.conv_out, nn.Sequential):
last_layer = self.decoder.conv_out[-1]
else:
last_layer = self.decoder.conv_out
else:
last_layer = self.decoder.layers[-1]
return last_layer
def set_use_tpu_flash_attention(self):
for block in self.decoder.up_blocks:
if isinstance(block, UNetMidBlock3D) and block.attention_blocks:
for attention_block in block.attention_blocks:
attention_block.set_use_tpu_flash_attention()
class Encoder(nn.Module):
r"""
The `Encoder` layer of a variational autoencoder that encodes its input into a latent representation.
Args:
dims (`int` or `Tuple[int, int]`, *optional*, defaults to 3):
The number of dimensions to use in convolutions.
in_channels (`int`, *optional*, defaults to 3):
The number of input channels.
out_channels (`int`, *optional*, defaults to 3):
The number of output channels.
blocks (`List[Tuple[str, int]]`, *optional*, defaults to `[("res_x", 1)]`):
The blocks to use. Each block is a tuple of the block name and the number of layers.
base_channels (`int`, *optional*, defaults to 128):
The number of output channels for the first convolutional layer.
norm_num_groups (`int`, *optional*, defaults to 32):
The number of groups for normalization.
patch_size (`int`, *optional*, defaults to 1):
The patch size to use. Should be a power of 2.
norm_layer (`str`, *optional*, defaults to `group_norm`):
The normalization layer to use. Can be either `group_norm` or `pixel_norm`.
latent_log_var (`str`, *optional*, defaults to `per_channel`):
The number of channels for the log variance. Can be either `per_channel`, `uniform`, `constant` or `none`.
"""
def __init__(
self,
dims: Union[int, Tuple[int, int]] = 3,
in_channels: int = 3,
out_channels: int = 3,
blocks: List[Tuple[str, int | dict]] = [("res_x", 1)],
base_channels: int = 128,
norm_num_groups: int = 32,
patch_size: Union[int, Tuple[int]] = 1,
norm_layer: str = "group_norm", # group_norm, pixel_norm
latent_log_var: str = "per_channel",
spatial_padding_mode: str = "zeros",
):
super().__init__()
self.patch_size = patch_size
self.norm_layer = norm_layer
self.latent_channels = out_channels
self.latent_log_var = latent_log_var
self.blocks_desc = blocks
in_channels = in_channels * patch_size**2
output_channel = base_channels
self.conv_in = make_conv_nd(
dims=dims,
in_channels=in_channels,
out_channels=output_channel,
kernel_size=3,
stride=1,
padding=1,
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
self.down_blocks = nn.ModuleList([])
for block_name, block_params in blocks:
input_channel = output_channel
if isinstance(block_params, int):
block_params = {"num_layers": block_params}
if block_name == "res_x":
block = UNetMidBlock3D(
dims=dims,
in_channels=input_channel,
num_layers=block_params["num_layers"],
resnet_eps=1e-6,
resnet_groups=norm_num_groups,
norm_layer=norm_layer,
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "res_x_y":
output_channel = block_params.get("multiplier", 2) * output_channel
block = ResnetBlock3D(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
eps=1e-6,
groups=norm_num_groups,
norm_layer=norm_layer,
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_time":
block = make_conv_nd(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
kernel_size=3,
stride=(2, 1, 1),
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_space":
block = make_conv_nd(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
kernel_size=3,
stride=(1, 2, 2),
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_all":
block = make_conv_nd(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
kernel_size=3,
stride=(2, 2, 2),
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_all_x_y":
output_channel = block_params.get("multiplier", 2) * output_channel
block = make_conv_nd(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
kernel_size=3,
stride=(2, 2, 2),
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_all_res":
output_channel = block_params.get("multiplier", 2) * output_channel
block = SpaceToDepthDownsample(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
stride=(2, 2, 2),
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_space_res":
output_channel = block_params.get("multiplier", 2) * output_channel
block = SpaceToDepthDownsample(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
stride=(1, 2, 2),
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_time_res":
output_channel = block_params.get("multiplier", 2) * output_channel
block = SpaceToDepthDownsample(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
stride=(2, 1, 1),
spatial_padding_mode=spatial_padding_mode,
)
else:
raise ValueError(f"unknown block: {block_name}")
self.down_blocks.append(block)
# out
if norm_layer == "group_norm":
self.conv_norm_out = nn.GroupNorm(
num_channels=output_channel, num_groups=norm_num_groups, eps=1e-6
)
elif norm_layer == "pixel_norm":
self.conv_norm_out = PixelNorm()
elif norm_layer == "layer_norm":
self.conv_norm_out = LayerNorm(output_channel, eps=1e-6)
self.conv_act = nn.SiLU()
conv_out_channels = out_channels
if latent_log_var == "per_channel":
conv_out_channels *= 2
elif latent_log_var == "uniform":
conv_out_channels += 1
elif latent_log_var == "constant":
conv_out_channels += 1
elif latent_log_var != "none":
raise ValueError(f"Invalid latent_log_var: {latent_log_var}")
self.conv_out = make_conv_nd(
dims,
output_channel,
conv_out_channels,
3,
padding=1,
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
self.gradient_checkpointing = False
def forward(self, sample: torch.FloatTensor) -> torch.FloatTensor:
r"""The forward method of the `Encoder` class."""
sample = patchify(sample, patch_size_hw=self.patch_size, patch_size_t=1)
sample = self.conv_in(sample)
checkpoint_fn = (
partial(torch.utils.checkpoint.checkpoint, use_reentrant=False)
if self.gradient_checkpointing and self.training
else lambda x: x
)
for down_block in self.down_blocks:
sample = checkpoint_fn(down_block)(sample)
sample = self.conv_norm_out(sample)
sample = self.conv_act(sample)
sample = self.conv_out(sample)
if self.latent_log_var == "uniform":
last_channel = sample[:, -1:, ...]
num_dims = sample.dim()
if num_dims == 4:
# For shape (B, C, H, W)
repeated_last_channel = last_channel.repeat(
1, sample.shape[1] - 2, 1, 1
)
sample = torch.cat([sample, repeated_last_channel], dim=1)
elif num_dims == 5:
# For shape (B, C, F, H, W)
repeated_last_channel = last_channel.repeat(
1, sample.shape[1] - 2, 1, 1, 1
)
sample = torch.cat([sample, repeated_last_channel], dim=1)
else:
raise ValueError(f"Invalid input shape: {sample.shape}")
elif self.latent_log_var == "constant":
sample = sample[:, :-1, ...]
approx_ln_0 = (
-30
) # this is the minimal clamp value in DiagonalGaussianDistribution objects
sample = torch.cat(
[sample, torch.ones_like(sample, device=sample.device) * approx_ln_0],
dim=1,
)
return sample
class Decoder(nn.Module):
r"""
The `Decoder` layer of a variational autoencoder that decodes its latent representation into an output sample.
Args:
dims (`int` or `Tuple[int, int]`, *optional*, defaults to 3):
The number of dimensions to use in convolutions.
in_channels (`int`, *optional*, defaults to 3):
The number of input channels.
out_channels (`int`, *optional*, defaults to 3):
The number of output channels.
blocks (`List[Tuple[str, int]]`, *optional*, defaults to `[("res_x", 1)]`):
The blocks to use. Each block is a tuple of the block name and the number of layers.
base_channels (`int`, *optional*, defaults to 128):
The number of output channels for the first convolutional layer.
norm_num_groups (`int`, *optional*, defaults to 32):
The number of groups for normalization.
patch_size (`int`, *optional*, defaults to 1):
The patch size to use. Should be a power of 2.
norm_layer (`str`, *optional*, defaults to `group_norm`):
The normalization layer to use. Can be either `group_norm` or `pixel_norm`.
causal (`bool`, *optional*, defaults to `True`):
Whether to use causal convolutions or not.
"""
def __init__(
self,
dims,
in_channels: int = 3,
out_channels: int = 3,
blocks: List[Tuple[str, int | dict]] = [("res_x", 1)],
base_channels: int = 128,
layers_per_block: int = 2,
norm_num_groups: int = 32,
patch_size: int = 1,
norm_layer: str = "group_norm",
causal: bool = True,
timestep_conditioning: bool = False,
spatial_padding_mode: str = "zeros",
):
super().__init__()
self.patch_size = patch_size
self.layers_per_block = layers_per_block
out_channels = out_channels * patch_size**2
self.causal = causal
self.blocks_desc = blocks
# Compute output channel to be product of all channel-multiplier blocks
output_channel = base_channels
for block_name, block_params in list(reversed(blocks)):
block_params = block_params if isinstance(block_params, dict) else {}
if block_name == "res_x_y":
output_channel = output_channel * block_params.get("multiplier", 2)
if block_name.startswith("compress"):
output_channel = output_channel * block_params.get("multiplier", 1)
self.conv_in = make_conv_nd(
dims,
in_channels,
output_channel,
kernel_size=3,
stride=1,
padding=1,
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
self.up_blocks = nn.ModuleList([])
for block_name, block_params in list(reversed(blocks)):
input_channel = output_channel
if isinstance(block_params, int):
block_params = {"num_layers": block_params}
if block_name == "res_x":
block = UNetMidBlock3D(
dims=dims,
in_channels=input_channel,
num_layers=block_params["num_layers"],
resnet_eps=1e-6,
resnet_groups=norm_num_groups,
norm_layer=norm_layer,
inject_noise=block_params.get("inject_noise", False),
timestep_conditioning=timestep_conditioning,
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "attn_res_x":
block = UNetMidBlock3D(
dims=dims,
in_channels=input_channel,
num_layers=block_params["num_layers"],
resnet_groups=norm_num_groups,
norm_layer=norm_layer,
inject_noise=block_params.get("inject_noise", False),
timestep_conditioning=timestep_conditioning,
attention_head_dim=block_params["attention_head_dim"],
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "res_x_y":
output_channel = output_channel // block_params.get("multiplier", 2)
block = ResnetBlock3D(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
eps=1e-6,
groups=norm_num_groups,
norm_layer=norm_layer,
inject_noise=block_params.get("inject_noise", False),
timestep_conditioning=False,
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_time":
block = DepthToSpaceUpsample(
dims=dims,
in_channels=input_channel,
stride=(2, 1, 1),
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_space":
block = DepthToSpaceUpsample(
dims=dims,
in_channels=input_channel,
stride=(1, 2, 2),
spatial_padding_mode=spatial_padding_mode,
)
elif block_name == "compress_all":
output_channel = output_channel // block_params.get("multiplier", 1)
block = DepthToSpaceUpsample(
dims=dims,
in_channels=input_channel,
stride=(2, 2, 2),
residual=block_params.get("residual", False),
out_channels_reduction_factor=block_params.get("multiplier", 1),
spatial_padding_mode=spatial_padding_mode,
)
else:
raise ValueError(f"unknown layer: {block_name}")
self.up_blocks.append(block)
if norm_layer == "group_norm":
self.conv_norm_out = nn.GroupNorm(
num_channels=output_channel, num_groups=norm_num_groups, eps=1e-6
)
elif norm_layer == "pixel_norm":
self.conv_norm_out = PixelNorm()
elif norm_layer == "layer_norm":
self.conv_norm_out = LayerNorm(output_channel, eps=1e-6)
self.conv_act = nn.SiLU()
self.conv_out = make_conv_nd(
dims,
output_channel,
out_channels,
3,
padding=1,
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
self.gradient_checkpointing = False
self.timestep_conditioning = timestep_conditioning
if timestep_conditioning:
self.timestep_scale_multiplier = nn.Parameter(
torch.tensor(1000.0, dtype=torch.float32)
)
self.last_time_embedder = PixArtAlphaCombinedTimestepSizeEmbeddings(
output_channel * 2, 0
)
self.last_scale_shift_table = nn.Parameter(
torch.randn(2, output_channel) / output_channel**0.5
)
def forward(
self,
sample: torch.FloatTensor,
target_shape,
timestep: Optional[torch.Tensor] = None,
) -> torch.FloatTensor:
r"""The forward method of the `Decoder` class."""
assert target_shape is not None, "target_shape must be provided"
batch_size = sample.shape[0]
sample = self.conv_in(sample, causal=self.causal)
upscale_dtype = next(iter(self.up_blocks.parameters())).dtype
checkpoint_fn = (
partial(torch.utils.checkpoint.checkpoint, use_reentrant=False)
if self.gradient_checkpointing and self.training
else lambda x: x
)
sample = sample.to(upscale_dtype)
if self.timestep_conditioning:
assert (
timestep is not None
), "should pass timestep with timestep_conditioning=True"
scaled_timestep = timestep * self.timestep_scale_multiplier
for up_block in self.up_blocks:
if self.timestep_conditioning and isinstance(up_block, UNetMidBlock3D):
sample = checkpoint_fn(up_block)(
sample, causal=self.causal, timestep=scaled_timestep
)
else:
sample = checkpoint_fn(up_block)(sample, causal=self.causal)
sample = self.conv_norm_out(sample)
if self.timestep_conditioning:
embedded_timestep = self.last_time_embedder(
timestep=scaled_timestep.flatten(),
resolution=None,
aspect_ratio=None,
batch_size=sample.shape[0],
hidden_dtype=sample.dtype,
)
embedded_timestep = embedded_timestep.view(
batch_size, embedded_timestep.shape[-1], 1, 1, 1
)
ada_values = self.last_scale_shift_table[
None, ..., None, None, None
] + embedded_timestep.reshape(
batch_size,
2,
-1,
embedded_timestep.shape[-3],
embedded_timestep.shape[-2],
embedded_timestep.shape[-1],
)
shift, scale = ada_values.unbind(dim=1)
sample = sample * (1 + scale) + shift
sample = self.conv_act(sample)
sample = self.conv_out(sample, causal=self.causal)
sample = unpatchify(sample, patch_size_hw=self.patch_size, patch_size_t=1)
return sample
class UNetMidBlock3D(nn.Module):
"""
A 3D UNet mid-block [`UNetMidBlock3D`] with multiple residual blocks.
Args:
in_channels (`int`): The number of input channels.
dropout (`float`, *optional*, defaults to 0.0): The dropout rate.
num_layers (`int`, *optional*, defaults to 1): The number of residual blocks.
resnet_eps (`float`, *optional*, 1e-6 ): The epsilon value for the resnet blocks.
resnet_groups (`int`, *optional*, defaults to 32):
The number of groups to use in the group normalization layers of the resnet blocks.
norm_layer (`str`, *optional*, defaults to `group_norm`):
The normalization layer to use. Can be either `group_norm` or `pixel_norm`.
inject_noise (`bool`, *optional*, defaults to `False`):
Whether to inject noise into the hidden states.
timestep_conditioning (`bool`, *optional*, defaults to `False`):
Whether to condition the hidden states on the timestep.
attention_head_dim (`int`, *optional*, defaults to -1):
The dimension of the attention head. If -1, no attention is used.
Returns:
`torch.FloatTensor`: The output of the last residual block, which is a tensor of shape `(batch_size,
in_channels, height, width)`.
"""
def __init__(
self,
dims: Union[int, Tuple[int, int]],
in_channels: int,
dropout: float = 0.0,
num_layers: int = 1,
resnet_eps: float = 1e-6,
resnet_groups: int = 32,
norm_layer: str = "group_norm",
inject_noise: bool = False,
timestep_conditioning: bool = False,
attention_head_dim: int = -1,
spatial_padding_mode: str = "zeros",
):
super().__init__()
resnet_groups = (
resnet_groups if resnet_groups is not None else min(in_channels // 4, 32)
)
self.timestep_conditioning = timestep_conditioning
if timestep_conditioning:
self.time_embedder = PixArtAlphaCombinedTimestepSizeEmbeddings(
in_channels * 4, 0
)
self.res_blocks = nn.ModuleList(
[
ResnetBlock3D(
dims=dims,
in_channels=in_channels,
out_channels=in_channels,
eps=resnet_eps,
groups=resnet_groups,
dropout=dropout,
norm_layer=norm_layer,
inject_noise=inject_noise,
timestep_conditioning=timestep_conditioning,
spatial_padding_mode=spatial_padding_mode,
)
for _ in range(num_layers)
]
)
self.attention_blocks = None
if attention_head_dim > 0:
if attention_head_dim > in_channels:
raise ValueError(
"attention_head_dim must be less than or equal to in_channels"
)
self.attention_blocks = nn.ModuleList(
[
Attention(
query_dim=in_channels,
heads=in_channels // attention_head_dim,
dim_head=attention_head_dim,
bias=True,
out_bias=True,
qk_norm="rms_norm",
residual_connection=True,
)
for _ in range(num_layers)
]
)
def forward(
self,
hidden_states: torch.FloatTensor,
causal: bool = True,
timestep: Optional[torch.Tensor] = None,
) -> torch.FloatTensor:
timestep_embed = None
if self.timestep_conditioning:
assert (
timestep is not None
), "should pass timestep with timestep_conditioning=True"
batch_size = hidden_states.shape[0]
timestep_embed = self.time_embedder(
timestep=timestep.flatten(),
resolution=None,
aspect_ratio=None,
batch_size=batch_size,
hidden_dtype=hidden_states.dtype,
)
timestep_embed = timestep_embed.view(
batch_size, timestep_embed.shape[-1], 1, 1, 1
)
if self.attention_blocks:
for resnet, attention in zip(self.res_blocks, self.attention_blocks):
hidden_states = resnet(
hidden_states, causal=causal, timestep=timestep_embed
)
# Reshape the hidden states to be (batch_size, frames * height * width, channel)
batch_size, channel, frames, height, width = hidden_states.shape
hidden_states = hidden_states.view(
batch_size, channel, frames * height * width
).transpose(1, 2)
if attention.use_tpu_flash_attention:
# Pad the second dimension to be divisible by block_k_major (block in flash attention)
seq_len = hidden_states.shape[1]
block_k_major = 512
pad_len = (block_k_major - seq_len % block_k_major) % block_k_major
if pad_len > 0:
hidden_states = F.pad(
hidden_states, (0, 0, 0, pad_len), "constant", 0
)
# Create a mask with ones for the original sequence length and zeros for the padded indexes
mask = torch.ones(
(hidden_states.shape[0], seq_len),
device=hidden_states.device,
dtype=hidden_states.dtype,
)
if pad_len > 0:
mask = F.pad(mask, (0, pad_len), "constant", 0)
hidden_states = attention(
hidden_states,
attention_mask=(
None if not attention.use_tpu_flash_attention else mask
),
)
if attention.use_tpu_flash_attention:
# Remove the padding
if pad_len > 0:
hidden_states = hidden_states[:, :-pad_len, :]
# Reshape the hidden states back to (batch_size, channel, frames, height, width, channel)
hidden_states = hidden_states.transpose(-1, -2).reshape(
batch_size, channel, frames, height, width
)
else:
for resnet in self.res_blocks:
hidden_states = resnet(
hidden_states, causal=causal, timestep=timestep_embed
)
return hidden_states
class SpaceToDepthDownsample(nn.Module):
def __init__(self, dims, in_channels, out_channels, stride, spatial_padding_mode):
super().__init__()
self.stride = stride
self.group_size = in_channels * np.prod(stride) // out_channels
self.conv = make_conv_nd(
dims=dims,
in_channels=in_channels,
out_channels=out_channels // np.prod(stride),
kernel_size=3,
stride=1,
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
def forward(self, x, causal: bool = True):
if self.stride[0] == 2:
x = torch.cat(
[x[:, :, :1, :, :], x], dim=2
) # duplicate first frames for padding
# skip connection
x_in = rearrange(
x,
"b c (d p1) (h p2) (w p3) -> b (c p1 p2 p3) d h w",
p1=self.stride[0],
p2=self.stride[1],
p3=self.stride[2],
)
x_in = rearrange(x_in, "b (c g) d h w -> b c g d h w", g=self.group_size)
x_in = x_in.mean(dim=2)
# conv
x = self.conv(x, causal=causal)
x = rearrange(
x,
"b c (d p1) (h p2) (w p3) -> b (c p1 p2 p3) d h w",
p1=self.stride[0],
p2=self.stride[1],
p3=self.stride[2],
)
x = x + x_in
return x
class DepthToSpaceUpsample(nn.Module):
def __init__(
self,
dims,
in_channels,
stride,
residual=False,
out_channels_reduction_factor=1,
spatial_padding_mode="zeros",
):
super().__init__()
self.stride = stride
self.out_channels = (
np.prod(stride) * in_channels // out_channels_reduction_factor
)
self.conv = make_conv_nd(
dims=dims,
in_channels=in_channels,
out_channels=self.out_channels,
kernel_size=3,
stride=1,
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
self.pixel_shuffle = PixelShuffleND(dims=dims, upscale_factors=stride)
self.residual = residual
self.out_channels_reduction_factor = out_channels_reduction_factor
def forward(self, x, causal: bool = True):
if self.residual:
# Reshape and duplicate the input to match the output shape
x_in = self.pixel_shuffle(x)
num_repeat = np.prod(self.stride) // self.out_channels_reduction_factor
x_in = x_in.repeat(1, num_repeat, 1, 1, 1)
if self.stride[0] == 2:
x_in = x_in[:, :, 1:, :, :]
x = self.conv(x, causal=causal)
x = self.pixel_shuffle(x)
if self.stride[0] == 2:
x = x[:, :, 1:, :, :]
if self.residual:
x = x + x_in
return x
class LayerNorm(nn.Module):
def __init__(self, dim, eps, elementwise_affine=True) -> None:
super().__init__()
self.norm = nn.LayerNorm(dim, eps=eps, elementwise_affine=elementwise_affine)
def forward(self, x):
x = rearrange(x, "b c d h w -> b d h w c")
x = self.norm(x)
x = rearrange(x, "b d h w c -> b c d h w")
return x
class ResnetBlock3D(nn.Module):
r"""
A Resnet block.
Parameters:
in_channels (`int`): The number of channels in the input.
out_channels (`int`, *optional*, default to be `None`):
The number of output channels for the first conv layer. If None, same as `in_channels`.
dropout (`float`, *optional*, defaults to `0.0`): The dropout probability to use.
groups (`int`, *optional*, default to `32`): The number of groups to use for the first normalization layer.
eps (`float`, *optional*, defaults to `1e-6`): The epsilon to use for the normalization.
"""
def __init__(
self,
dims: Union[int, Tuple[int, int]],
in_channels: int,
out_channels: Optional[int] = None,
dropout: float = 0.0,
groups: int = 32,
eps: float = 1e-6,
norm_layer: str = "group_norm",
inject_noise: bool = False,
timestep_conditioning: bool = False,
spatial_padding_mode: str = "zeros",
):
super().__init__()
self.in_channels = in_channels
out_channels = in_channels if out_channels is None else out_channels
self.out_channels = out_channels
self.inject_noise = inject_noise
if norm_layer == "group_norm":
self.norm1 = nn.GroupNorm(
num_groups=groups, num_channels=in_channels, eps=eps, affine=True
)
elif norm_layer == "pixel_norm":
self.norm1 = PixelNorm()
elif norm_layer == "layer_norm":
self.norm1 = LayerNorm(in_channels, eps=eps, elementwise_affine=True)
self.non_linearity = nn.SiLU()
self.conv1 = make_conv_nd(
dims,
in_channels,
out_channels,
kernel_size=3,
stride=1,
padding=1,
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
if inject_noise:
self.per_channel_scale1 = nn.Parameter(torch.zeros((in_channels, 1, 1)))
if norm_layer == "group_norm":
self.norm2 = nn.GroupNorm(
num_groups=groups, num_channels=out_channels, eps=eps, affine=True
)
elif norm_layer == "pixel_norm":
self.norm2 = PixelNorm()
elif norm_layer == "layer_norm":
self.norm2 = LayerNorm(out_channels, eps=eps, elementwise_affine=True)
self.dropout = torch.nn.Dropout(dropout)
self.conv2 = make_conv_nd(
dims,
out_channels,
out_channels,
kernel_size=3,
stride=1,
padding=1,
causal=True,
spatial_padding_mode=spatial_padding_mode,
)
if inject_noise:
self.per_channel_scale2 = nn.Parameter(torch.zeros((in_channels, 1, 1)))
self.conv_shortcut = (
make_linear_nd(
dims=dims, in_channels=in_channels, out_channels=out_channels
)
if in_channels != out_channels
else nn.Identity()
)
self.norm3 = (
LayerNorm(in_channels, eps=eps, elementwise_affine=True)
if in_channels != out_channels
else nn.Identity()
)
self.timestep_conditioning = timestep_conditioning
if timestep_conditioning:
self.scale_shift_table = nn.Parameter(
torch.randn(4, in_channels) / in_channels**0.5
)
def _feed_spatial_noise(
self, hidden_states: torch.FloatTensor, per_channel_scale: torch.FloatTensor
) -> torch.FloatTensor:
spatial_shape = hidden_states.shape[-2:]
device = hidden_states.device
dtype = hidden_states.dtype
# similar to the "explicit noise inputs" method in style-gan
spatial_noise = torch.randn(spatial_shape, device=device, dtype=dtype)[None]
scaled_noise = (spatial_noise * per_channel_scale)[None, :, None, ...]
hidden_states = hidden_states + scaled_noise
return hidden_states
def forward(
self,
input_tensor: torch.FloatTensor,
causal: bool = True,
timestep: Optional[torch.Tensor] = None,
) -> torch.FloatTensor:
hidden_states = input_tensor
batch_size = hidden_states.shape[0]
hidden_states = self.norm1(hidden_states)
if self.timestep_conditioning:
assert (
timestep is not None
), "should pass timestep with timestep_conditioning=True"
ada_values = self.scale_shift_table[
None, ..., None, None, None
] + timestep.reshape(
batch_size,
4,
-1,
timestep.shape[-3],
timestep.shape[-2],
timestep.shape[-1],
)
shift1, scale1, shift2, scale2 = ada_values.unbind(dim=1)
hidden_states = hidden_states * (1 + scale1) + shift1
hidden_states = self.non_linearity(hidden_states)
hidden_states = self.conv1(hidden_states, causal=causal)
if self.inject_noise:
hidden_states = self._feed_spatial_noise(
hidden_states, self.per_channel_scale1
)
hidden_states = self.norm2(hidden_states)
if self.timestep_conditioning:
hidden_states = hidden_states * (1 + scale2) + shift2
hidden_states = self.non_linearity(hidden_states)
hidden_states = self.dropout(hidden_states)
hidden_states = self.conv2(hidden_states, causal=causal)
if self.inject_noise:
hidden_states = self._feed_spatial_noise(
hidden_states, self.per_channel_scale2
)
input_tensor = self.norm3(input_tensor)
batch_size = input_tensor.shape[0]
input_tensor = self.conv_shortcut(input_tensor)
output_tensor = input_tensor + hidden_states
return output_tensor
def patchify(x, patch_size_hw, patch_size_t=1):
if patch_size_hw == 1 and patch_size_t == 1:
return x
if x.dim() == 4:
x = rearrange(
x, "b c (h q) (w r) -> b (c r q) h w", q=patch_size_hw, r=patch_size_hw
)
elif x.dim() == 5:
x = rearrange(
x,
"b c (f p) (h q) (w r) -> b (c p r q) f h w",
p=patch_size_t,
q=patch_size_hw,
r=patch_size_hw,
)
else:
raise ValueError(f"Invalid input shape: {x.shape}")
return x
def unpatchify(x, patch_size_hw, patch_size_t=1):
if patch_size_hw == 1 and patch_size_t == 1:
return x
if x.dim() == 4:
x = rearrange(
x, "b (c r q) h w -> b c (h q) (w r)", q=patch_size_hw, r=patch_size_hw
)
elif x.dim() == 5:
x = rearrange(
x,
"b (c p r q) f h w -> b c (f p) (h q) (w r)",
p=patch_size_t,
q=patch_size_hw,
r=patch_size_hw,
)
return x
def create_video_autoencoder_demo_config(
latent_channels: int = 64,
):
encoder_blocks = [
("res_x", {"num_layers": 2}),
("compress_space_res", {"multiplier": 2}),
("compress_time_res", {"multiplier": 2}),
("compress_all_res", {"multiplier": 2}),
("compress_all_res", {"multiplier": 2}),
("res_x", {"num_layers": 1}),
]
decoder_blocks = [
("res_x", {"num_layers": 2, "inject_noise": False}),
("compress_all", {"residual": True, "multiplier": 2}),
("compress_all", {"residual": True, "multiplier": 2}),
("compress_all", {"residual": True, "multiplier": 2}),
("res_x", {"num_layers": 2, "inject_noise": False}),
]
return {
"_class_name": "CausalVideoAutoencoder",
"dims": 3,
"encoder_blocks": encoder_blocks,
"decoder_blocks": decoder_blocks,
"latent_channels": latent_channels,
"norm_layer": "pixel_norm",
"patch_size": 4,
"latent_log_var": "uniform",
"use_quant_conv": False,
"causal_decoder": False,
"timestep_conditioning": True,
"spatial_padding_mode": "replicate",
}
def test_vae_patchify_unpatchify():
import torch
x = torch.randn(2, 3, 8, 64, 64)
x_patched = patchify(x, patch_size_hw=4, patch_size_t=4)
x_unpatched = unpatchify(x_patched, patch_size_hw=4, patch_size_t=4)
assert torch.allclose(x, x_unpatched)
def demo_video_autoencoder_forward_backward():
# Configuration for the VideoAutoencoder
config = create_video_autoencoder_demo_config()
# Instantiate the VideoAutoencoder with the specified configuration
video_autoencoder = CausalVideoAutoencoder.from_config(config)
print(video_autoencoder)
video_autoencoder.eval()
# Print the total number of parameters in the video autoencoder
total_params = sum(p.numel() for p in video_autoencoder.parameters())
print(f"Total number of parameters in VideoAutoencoder: {total_params:,}")
# Create a mock input tensor simulating a batch of videos
# Shape: (batch_size, channels, depth, height, width)
# E.g., 4 videos, each with 3 color channels, 16 frames, and 64x64 pixels per frame
input_videos = torch.randn(2, 3, 17, 64, 64)
# Forward pass: encode and decode the input videos
latent = video_autoencoder.encode(input_videos).latent_dist.mode()
print(f"input shape={input_videos.shape}")
print(f"latent shape={latent.shape}")
timestep = torch.ones(input_videos.shape[0]) * 0.1
reconstructed_videos = video_autoencoder.decode(
latent, target_shape=input_videos.shape, timestep=timestep
).sample
print(f"reconstructed shape={reconstructed_videos.shape}")
# Validate that single image gets treated the same way as first frame
input_image = input_videos[:, :, :1, :, :]
image_latent = video_autoencoder.encode(input_image).latent_dist.mode()
_ = video_autoencoder.decode(
image_latent, target_shape=image_latent.shape, timestep=timestep
).sample
first_frame_latent = latent[:, :, :1, :, :]
assert torch.allclose(image_latent, first_frame_latent, atol=1e-6)
# assert torch.allclose(reconstructed_image, reconstructed_videos[:, :, :1, :, :], atol=1e-6)
# assert torch.allclose(image_latent, first_frame_latent, atol=1e-6)
# assert (reconstructed_image == reconstructed_videos[:, :, :1, :, :]).all()
# Calculate the loss (e.g., mean squared error)
loss = torch.nn.functional.mse_loss(input_videos, reconstructed_videos)
# Perform backward pass
loss.backward()
print(f"Demo completed with loss: {loss.item()}")
# Ensure to call the demo function to execute the forward and backward pass
if __name__ == "__main__":
demo_video_autoencoder_forward_backward()
================================================
FILE: ltx_video/models/autoencoders/conv_nd_factory.py
================================================
from typing import Tuple, Union
import torch
from ltx_video.models.autoencoders.dual_conv3d import DualConv3d
from ltx_video.models.autoencoders.causal_conv3d import CausalConv3d
def make_conv_nd(
dims: Union[int, Tuple[int, int]],
in_channels: int,
out_channels: int,
kernel_size: int,
stride=1,
padding=0,
dilation=1,
groups=1,
bias=True,
causal=False,
spatial_padding_mode="zeros",
temporal_padding_mode="zeros",
):
if not (spatial_padding_mode == temporal_padding_mode or causal):
raise NotImplementedError("spatial and temporal padding modes must be equal")
if dims == 2:
return torch.nn.Conv2d(
in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
dilation=dilation,
groups=groups,
bias=bias,
padding_mode=spatial_padding_mode,
)
elif dims == 3:
if causal:
return CausalConv3d(
in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
dilation=dilation,
groups=groups,
bias=bias,
spatial_padding_mode=spatial_padding_mode,
)
return torch.nn.Conv3d(
in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
dilation=dilation,
groups=groups,
bias=bias,
padding_mode=spatial_padding_mode,
)
elif dims == (2, 1):
return DualConv3d(
in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
bias=bias,
padding_mode=spatial_padding_mode,
)
else:
raise ValueError(f"unsupported dimensions: {dims}")
def make_linear_nd(
dims: int,
in_channels: int,
out_channels: int,
bias=True,
):
if dims == 2:
return torch.nn.Conv2d(
in_channels=in_channels, out_channels=out_channels, kernel_size=1, bias=bias
)
elif dims == 3 or dims == (2, 1):
return torch.nn.Conv3d(
in_channels=in_channels, out_channels=out_channels, kernel_size=1, bias=bias
)
else:
raise ValueError(f"unsupported dimensions: {dims}")
================================================
FILE: ltx_video/models/autoencoders/dual_conv3d.py
================================================
import math
from typing import Tuple, Union
import torch
import torch.nn as nn
import torch.nn.functional as F
from einops import rearrange
class DualConv3d(nn.Module):
def __init__(
self,
in_channels,
out_channels,
kernel_size,
stride: Union[int, Tuple[int, int, int]] = 1,
padding: Union[int, Tuple[int, int, int]] = 0,
dilation: Union[int, Tuple[int, int, int]] = 1,
groups=1,
bias=True,
padding_mode="zeros",
):
super(DualConv3d, self).__init__()
self.in_channels = in_channels
self.out_channels = out_channels
self.padding_mode = padding_mode
# Ensure kernel_size, stride, padding, and dilation are tuples of length 3
if isinstance(kernel_size, int):
kernel_size = (kernel_size, kernel_size, kernel_size)
if kernel_size == (1, 1, 1):
raise ValueError(
"kernel_size must be greater than 1. Use make_linear_nd instead."
)
if isinstance(stride, int):
stride = (stride, stride, stride)
if isinstance(padding, int):
padding = (padding, padding, padding)
if isinstance(dilation, int):
dilation = (dilation, dilation, dilation)
# Set parameters for convolutions
self.groups = groups
self.bias = bias
# Define the size of the channels after the first convolution
intermediate_channels = (
out_channels if in_channels < out_channels else in_channels
)
# Define parameters for the first convolution
self.weight1 = nn.Parameter(
torch.Tensor(
intermediate_channels,
in_channels // groups,
1,
kernel_size[1],
kernel_size[2],
)
)
self.stride1 = (1, stride[1], stride[2])
self.padding1 = (0, padding[1], padding[2])
self.dilation1 = (1, dilation[1], dilation[2])
if bias:
self.bias1 = nn.Parameter(torch.Tensor(intermediate_channels))
else:
self.register_parameter("bias1", None)
# Define parameters for the second convolution
self.weight2 = nn.Parameter(
torch.Tensor(
out_channels, intermediate_channels // groups, kernel_size[0], 1, 1
)
)
self.stride2 = (stride[0], 1, 1)
self.padding2 = (padding[0], 0, 0)
self.dilation2 = (dilation[0], 1, 1)
if bias:
self.bias2 = nn.Parameter(torch.Tensor(out_channels))
else:
self.register_parameter("bias2", None)
# Initialize weights and biases
self.reset_parameters()
def reset_parameters(self):
nn.init.kaiming_uniform_(self.weight1, a=math.sqrt(5))
nn.init.kaiming_uniform_(self.weight2, a=math.sqrt(5))
if self.bias:
fan_in1, _ = nn.init._calculate_fan_in_and_fan_out(self.weight1)
bound1 = 1 / math.sqrt(fan_in1)
nn.init.uniform_(self.bias1, -bound1, bound1)
fan_in2, _ = nn.init._calculate_fan_in_and_fan_out(self.weight2)
bound2 = 1 / math.sqrt(fan_in2)
nn.init.uniform_(self.bias2, -bound2, bound2)
def forward(self, x, use_conv3d=False, skip_time_conv=False):
if use_conv3d:
return self.forward_with_3d(x=x, skip_time_conv=skip_time_conv)
else:
return self.forward_with_2d(x=x, skip_time_conv=skip_time_conv)
def forward_with_3d(self, x, skip_time_conv):
# First convolution
x = F.conv3d(
x,
self.weight1,
self.bias1,
self.stride1,
self.padding1,
self.dilation1,
self.groups,
padding_mode=self.padding_mode,
)
if skip_time_conv:
return x
# Second convolution
x = F.conv3d(
x,
self.weight2,
self.bias2,
self.stride2,
self.padding2,
self.dilation2,
self.groups,
padding_mode=self.padding_mode,
)
return x
def forward_with_2d(self, x, skip_time_conv):
b, c, d, h, w = x.shape
# First 2D convolution
x = rearrange(x, "b c d h w -> (b d) c h w")
# Squeeze the depth dimension out of weight1 since it's 1
weight1 = self.weight1.squeeze(2)
# Select stride, padding, and dilation for the 2D convolution
stride1 = (self.stride1[1], self.stride1[2])
padding1 = (self.padding1[1], self.padding1[2])
dilation1 = (self.dilation1[1], self.dilation1[2])
x = F.conv2d(
x,
weight1,
self.bias1,
stride1,
padding1,
dilation1,
self.groups,
padding_mode=self.padding_mode,
)
_, _, h, w = x.shape
if skip_time_conv:
x = rearrange(x, "(b d) c h w -> b c d h w", b=b)
return x
# Second convolution which is essentially treated as a 1D convolution across the 'd' dimension
x = rearrange(x, "(b d) c h w -> (b h w) c d", b=b)
# Reshape weight2 to match the expected dimensions for conv1d
weight2 = self.weight2.squeeze(-1).squeeze(-1)
# Use only the relevant dimension for stride, padding, and dilation for the 1D convolution
stride2 = self.stride2[0]
padding2 = self.padding2[0]
dilation2 = self.dilation2[0]
x = F.conv1d(
x,
weight2,
self.bias2,
stride2,
padding2,
dilation2,
self.groups,
padding_mode=self.padding_mode,
)
x = rearrange(x, "(b h w) c d -> b c d h w", b=b, h=h, w=w)
return x
@property
def weight(self):
return self.weight2
def test_dual_conv3d_consistency():
# Initialize parameters
in_channels = 3
out_channels = 5
kernel_size = (3, 3, 3)
stride = (2, 2, 2)
padding = (1, 1, 1)
# Create an instance of the DualConv3d class
dual_conv3d = DualConv3d(
in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
bias=True,
)
# Example input tensor
test_input = torch.randn(1, 3, 10, 10, 10)
# Perform forward passes with both 3D and 2D settings
output_conv3d = dual_conv3d(test_input, use_conv3d=True)
output_2d = dual_conv3d(test_input, use_conv3d=False)
# Assert that the outputs from both methods are sufficiently close
assert torch.allclose(
output_conv3d, output_2d, atol=1e-6
), "Outputs are not consistent between 3D and 2D convolutions."
================================================
FILE: ltx_video/models/autoencoders/latent_upsampler.py
================================================
from typing import Optional, Union
from pathlib import Path
import os
import json
import torch
import torch.nn as nn
from einops import rearrange
from diffusers import ConfigMixin, ModelMixin
from safetensors.torch import safe_open
from ltx_video.models.autoencoders.pixel_shuffle import PixelShuffleND
class ResBlock(nn.Module):
def __init__(
self, channels: int, mid_channels: Optional[int] = None, dims: int = 3
):
super().__init__()
if mid_channels is None:
mid_channels = channels
Conv = nn.Conv2d if dims == 2 else nn.Conv3d
self.conv1 = Conv(channels, mid_channels, kernel_size=3, padding=1)
self.norm1 = nn.GroupNorm(32, mid_channels)
self.conv2 = Conv(mid_channels, channels, kernel_size=3, padding=1)
self.norm2 = nn.GroupNorm(32, channels)
self.activation = nn.SiLU()
def forward(self, x: torch.Tensor) -> torch.Tensor:
residual = x
x = self.conv1(x)
x = self.norm1(x)
x = self.activation(x)
x = self.conv2(x)
x = self.norm2(x)
x = self.activation(x + residual)
return x
class LatentUpsampler(ModelMixin, ConfigMixin):
"""
Model to spatially upsample VAE latents.
Args:
in_channels (`int`): Number of channels in the input latent
mid_channels (`int`): Number of channels in the middle layers
num_blocks_per_stage (`int`): Number of ResBlocks to use in each stage (pre/post upsampling)
dims (`int`): Number of dimensions for convolutions (2 or 3)
spatial_upsample (`bool`): Whether to spatially upsample the latent
temporal_upsample (`bool`): Whether to temporally upsample the latent
"""
def __init__(
self,
in_channels: int = 128,
mid_channels: int = 512,
num_blocks_per_stage: int = 4,
dims: int = 3,
spatial_upsample: bool = True,
temporal_upsample: bool = False,
):
super().__init__()
self.in_channels = in_channels
self.mid_channels = mid_channels
self.num_blocks_per_stage = num_blocks_per_stage
self.dims = dims
self.spatial_upsample = spatial_upsample
self.temporal_upsample = temporal_upsample
Conv = nn.Conv2d if dims == 2 else nn.Conv3d
self.initial_conv = Conv(in_channels, mid_channels, kernel_size=3, padding=1)
self.initial_norm = nn.GroupNorm(32, mid_channels)
self.initial_activation = nn.SiLU()
self.res_blocks = nn.ModuleList(
[ResBlock(mid_channels, dims=dims) for _ in range(num_blocks_per_stage)]
)
if spatial_upsample and temporal_upsample:
self.upsampler = nn.Sequential(
nn.Conv3d(mid_channels, 8 * mid_channels, kernel_size=3, padding=1),
PixelShuffleND(3),
)
elif spatial_upsample:
self.upsampler = nn.Sequential(
nn.Conv2d(mid_channels, 4 * mid_channels, kernel_size=3, padding=1),
PixelShuffleND(2),
)
elif temporal_upsample:
self.upsampler = nn.Sequential(
nn.Conv3d(mid_channels, 2 * mid_channels, kernel_size=3, padding=1),
PixelShuffleND(1),
)
else:
raise ValueError(
"Either spatial_upsample or temporal_upsample must be True"
)
self.post_upsample_res_blocks = nn.ModuleList(
[ResBlock(mid_channels, dims=dims) for _ in range(num_blocks_per_stage)]
)
self.final_conv = Conv(mid_channels, in_channels, kernel_size=3, padding=1)
def forward(self, latent: torch.Tensor) -> torch.Tensor:
b, c, f, h, w = latent.shape
if self.dims == 2:
x = rearrange(latent, "b c f h w -> (b f) c h w")
x = self.initial_conv(x)
x = self.initial_norm(x)
x = self.initial_activation(x)
for block in self.res_blocks:
x = block(x)
x = self.upsampler(x)
for block in self.post_upsample_res_blocks:
x = block(x)
x = self.final_conv(x)
x = rearrange(x, "(b f) c h w -> b c f h w", b=b, f=f)
else:
x = self.initial_conv(latent)
x = self.initial_norm(x)
x = self.initial_activation(x)
for block in self.res_blocks:
x = block(x)
if self.temporal_upsample:
x = self.upsampler(x)
x = x[:, :, 1:, :, :]
else:
x = rearrange(x, "b c f h w -> (b f) c h w")
x = self.upsampler(x)
x = rearrange(x, "(b f) c h w -> b c f h w", b=b, f=f)
for block in self.post_upsample_res_blocks:
x = block(x)
x = self.final_conv(x)
return x
@classmethod
def from_config(cls, config):
return cls(
in_channels=config.get("in_channels", 4),
mid_channels=config.get("mid_channels", 128),
num_blocks_per_stage=config.get("num_blocks_per_stage", 4),
dims=config.get("dims", 2),
spatial_upsample=config.get("spatial_upsample", True),
temporal_upsample=config.get("temporal_upsample", False),
)
def config(self):
return {
"_class_name": "LatentUpsampler",
"in_channels": self.in_channels,
"mid_channels": self.mid_channels,
"num_blocks_per_stage": self.num_blocks_per_stage,
"dims": self.dims,
"spatial_upsample": self.spatial_upsample,
"temporal_upsample": self.temporal_upsample,
}
@classmethod
def from_pretrained(
cls,
pretrained_model_path: Optional[Union[str, os.PathLike]],
*args,
**kwargs,
):
pretrained_model_path = Path(pretrained_model_path)
if pretrained_model_path.is_file() and str(pretrained_model_path).endswith(
".safetensors"
):
state_dict = {}
with safe_open(pretrained_model_path, framework="pt", device="cpu") as f:
metadata = f.metadata()
for k in f.keys():
state_dict[k] = f.get_tensor(k)
config = json.loads(metadata["config"])
with torch.device("meta"):
latent_upsampler = LatentUpsampler.from_config(config)
latent_upsampler.load_state_dict(state_dict, assign=True)
return latent_upsampler
if __name__ == "__main__":
latent_upsampler = LatentUpsampler(num_blocks_per_stage=4, dims=3)
print(latent_upsampler)
total_params = sum(p.numel() for p in latent_upsampler.parameters())
print(f"Total number of parameters: {total_params:,}")
latent = torch.randn(1, 128, 9, 16, 16)
upsampled_latent = latent_upsampler(latent)
print(f"Upsampled latent shape: {upsampled_latent.shape}")
================================================
FILE: ltx_video/models/autoencoders/pixel_norm.py
================================================
import torch
from torch import nn
class PixelNorm(nn.Module):
def __init__(self, dim=1, eps=1e-8):
super(PixelNorm, self).__init__()
self.dim = dim
self.eps = eps
def forward(self, x):
return x / torch.sqrt(torch.mean(x**2, dim=self.dim, keepdim=True) + self.eps)
================================================
FILE: ltx_video/models/autoencoders/pixel_shuffle.py
================================================
import torch.nn as nn
from einops import rearrange
class PixelShuffleND(nn.Module):
def __init__(self, dims, upscale_factors=(2, 2, 2)):
super().__init__()
assert dims in [1, 2, 3], "dims must be 1, 2, or 3"
self.dims = dims
self.upscale_factors = upscale_factors
def forward(self, x):
if self.dims == 3:
return rearrange(
x,
"b (c p1 p2 p3) d h w -> b c (d p1) (h p2) (w p3)",
p1=self.upscale_factors[0],
p2=self.upscale_factors[1],
p3=self.upscale_factors[2],
)
elif self.dims == 2:
return rearrange(
x,
"b (c p1 p2) h w -> b c (h p1) (w p2)",
p1=self.upscale_factors[0],
p2=self.upscale_factors[1],
)
elif self.dims == 1:
return rearrange(
x,
"b (c p1) f h w -> b c (f p1) h w",
p1=self.upscale_factors[0],
)
================================================
FILE: ltx_video/models/autoencoders/vae.py
================================================
from typing import Optional, Union
import torch
import inspect
import math
import torch.nn as nn
from diffusers import ConfigMixin, ModelMixin
from diffusers.models.autoencoders.vae import (
DecoderOutput,
DiagonalGaussianDistribution,
)
from diffusers.models.modeling_outputs import AutoencoderKLOutput
from ltx_video.models.autoencoders.conv_nd_factory import make_conv_nd
class AutoencoderKLWrapper(ModelMixin, ConfigMixin):
"""Variational Autoencoder (VAE) model with KL loss.
VAE from the paper Auto-Encoding Variational Bayes by Diederik P. Kingma and Max Welling.
This model is a wrapper around an encoder and a decoder, and it adds a KL loss term to the reconstruction loss.
Args:
encoder (`nn.Module`):
Encoder module.
decoder (`nn.Module`):
Decoder module.
latent_channels (`int`, *optional*, defaults to 4):
Number of latent channels.
"""
def __init__(
self,
encoder: nn.Module,
decoder: nn.Module,
latent_channels: int = 4,
dims: int = 2,
sample_size=512,
use_quant_conv: bool = True,
normalize_latent_channels: bool = False,
):
super().__init__()
# pass init params to Encoder
self.encoder = encoder
self.use_quant_conv = use_quant_conv
self.normalize_latent_channels = normalize_latent_channels
# pass init params to Decoder
quant_dims = 2 if dims == 2 else 3
self.decoder = decoder
if use_quant_conv:
self.quant_conv = make_conv_nd(
quant_dims, 2 * latent_channels, 2 * latent_channels, 1
)
self.post_quant_conv = make_conv_nd(
quant_dims, latent_channels, latent_channels, 1
)
else:
self.quant_conv = nn.Identity()
self.post_quant_conv = nn.Identity()
if normalize_latent_channels:
if dims == 2:
self.latent_norm_out = nn.BatchNorm2d(latent_channels, affine=False)
else:
self.latent_norm_out = nn.BatchNorm3d(latent_channels, affine=False)
else:
self.latent_norm_out = nn.Identity()
self.use_z_tiling = False
self.use_hw_tiling = False
self.dims = dims
self.z_sample_size = 1
self.decoder_params = inspect.signature(self.decoder.forward).parameters
# only relevant if vae tiling is enabled
self.set_tiling_params(sample_size=sample_size, overlap_factor=0.25)
def set_tiling_params(self, sample_size: int = 512, overlap_factor: float = 0.25):
self.tile_sample_min_size = sample_size
num_blocks = len(self.encoder.down_blocks)
self.tile_latent_min_size = int(sample_size / (2 ** (num_blocks - 1)))
self.tile_overlap_factor = overlap_factor
def enable_z_tiling(self, z_sample_size: int = 8):
r"""
Enable tiling during VAE decoding.
When this option is enabled, the VAE will split the input tensor in tiles to compute decoding in several
steps. This is useful to save some memory and allow larger batch sizes.
"""
self.use_z_tiling = z_sample_size > 1
self.z_sample_size = z_sample_size
assert (
z_sample_size % 8 == 0 or z_sample_size == 1
), f"z_sample_size must be a multiple of 8 or 1. Got {z_sample_size}."
def disable_z_tiling(self):
r"""
Disable tiling during VAE decoding. If `use_tiling` was previously invoked, this method will go back to computing
decoding in one step.
"""
self.use_z_tiling = False
def enable_hw_tiling(self):
r"""
Enable tiling during VAE decoding along the height and width dimension.
"""
self.use_hw_tiling = True
def disable_hw_tiling(self):
r"""
Disable tiling during VAE decoding along the height and width dimension.
"""
self.use_hw_tiling = False
def _hw_tiled_encode(self, x: torch.FloatTensor, return_dict: bool = True):
overlap_size = int(self.tile_sample_min_size * (1 - self.tile_overlap_factor))
blend_extent = int(self.tile_latent_min_size * self.tile_overlap_factor)
row_limit = self.tile_latent_min_size - blend_extent
# Split the image into 512x512 tiles and encode them separately.
rows = []
for i in range(0, x.shape[3], overlap_size):
row = []
for j in range(0, x.shape[4], overlap_size):
tile = x[
:,
:,
:,
i : i + self.tile_sample_min_size,
j : j + self.tile_sample_min_size,
]
tile = self.encoder(tile)
tile = self.quant_conv(tile)
row.append(tile)
rows.append(row)
result_rows = []
for i, row in enumerate(rows):
result_row = []
for j, tile in enumerate(row):
# blend the above tile and the left tile
# to the current tile and add the current tile to the result row
if i > 0:
tile = self.blend_v(rows[i - 1][j], tile, blend_extent)
if j > 0:
tile = self.blend_h(row[j - 1], tile, blend_extent)
result_row.append(tile[:, :, :, :row_limit, :row_limit])
result_rows.append(torch.cat(result_row, dim=4))
moments = torch.cat(result_rows, dim=3)
return moments
def blend_z(
self, a: torch.Tensor, b: torch.Tensor, blend_extent: int
) -> torch.Tensor:
blend_extent = min(a.shape[2], b.shape[2], blend_extent)
for z in range(blend_extent):
b[:, :, z, :, :] = a[:, :, -blend_extent + z, :, :] * (
1 - z / blend_extent
) + b[:, :, z, :, :] * (z / blend_extent)
return b
def blend_v(
self, a: torch.Tensor, b: torch.Tensor, blend_extent: int
) -> torch.Tensor:
blend_extent = min(a.shape[3], b.shape[3], blend_extent)
for y in range(blend_extent):
b[:, :, :, y, :] = a[:, :, :, -blend_extent + y, :] * (
1 - y / blend_extent
) + b[:, :, :, y, :] * (y / blend_extent)
return b
def blend_h(
self, a: torch.Tensor, b: torch.Tensor, blend_extent: int
) -> torch.Tensor:
blend_extent = min(a.shape[4], b.shape[4], blend_extent)
for x in range(blend_extent):
b[:, :, :, :, x] = a[:, :, :, :, -blend_extent + x] * (
1 - x / blend_extent
) + b[:, :, :, :, x] * (x / blend_extent)
return b
def _hw_tiled_decode(self, z: torch.FloatTensor, target_shape):
overlap_size = int(self.tile_latent_min_size * (1 - self.tile_overlap_factor))
blend_extent = int(self.tile_sample_min_size * self.tile_overlap_factor)
row_limit = self.tile_sample_min_size - blend_extent
tile_target_shape = (
*target_shape[:3],
self.tile_sample_min_size,
self.tile_sample_min_size,
)
# Split z into overlapping 64x64 tiles and decode them separately.
# The tiles have an overlap to avoid seams between tiles.
rows = []
for i in range(0, z.shape[3], overlap_size):
row = []
for j in range(0, z.shape[4], overlap_size):
tile = z[
:,
:,
:,
i : i + self.tile_latent_min_size,
j : j + self.tile_latent_min_size,
]
tile = self.post_quant_conv(tile)
decoded = self.decoder(tile, target_shape=tile_target_shape)
row.append(decoded)
rows.append(row)
result_rows = []
for i, row in enumerate(rows):
result_row = []
for j, tile in enumerate(row):
# blend the above tile and the left tile
# to the current tile and add the current tile to the result row
if i > 0:
tile = self.blend_v(rows[i - 1][j], tile, blend_extent)
if j > 0:
tile = self.blend_h(row[j - 1], tile, blend_extent)
result_row.append(tile[:, :, :, :row_limit, :row_limit])
result_rows.append(torch.cat(result_row, dim=4))
dec = torch.cat(result_rows, dim=3)
return dec
def encode(
self, z: torch.FloatTensor, return_dict: bool = True
) -> Union[DecoderOutput, torch.FloatTensor]:
if self.use_z_tiling and z.shape[2] > self.z_sample_size > 1:
num_splits = z.shape[2] // self.z_sample_size
sizes = [self.z_sample_size] * num_splits
sizes = (
sizes + [z.shape[2] - sum(sizes)]
if z.shape[2] - sum(sizes) > 0
else sizes
)
tiles = z.split(sizes, dim=2)
moments_tiles = [
(
self._hw_tiled_encode(z_tile, return_dict)
if self.use_hw_tiling
else self._encode(z_tile)
)
for z_tile in tiles
]
moments = torch.cat(moments_tiles, dim=2)
else:
moments = (
self._hw_tiled_encode(z, return_dict)
if self.use_hw_tiling
else self._encode(z)
)
posterior = DiagonalGaussianDistribution(moments)
if not return_dict:
return (posterior,)
return AutoencoderKLOutput(latent_dist=posterior)
def _normalize_latent_channels(self, z: torch.FloatTensor) -> torch.FloatTensor:
if isinstance(self.latent_norm_out, nn.BatchNorm3d):
_, c, _, _, _ = z.shape
z = torch.cat(
[
self.latent_norm_out(z[:, : c // 2, :, :, :]),
z[:, c // 2 :, :, :, :],
],
dim=1,
)
elif isinstance(self.latent_norm_out, nn.BatchNorm2d):
raise NotImplementedError("BatchNorm2d not supported")
return z
def _unnormalize_latent_channels(self, z: torch.FloatTensor) -> torch.FloatTensor:
if isinstance(self.latent_norm_out, nn.BatchNorm3d):
running_mean = self.latent_norm_out.running_mean.view(1, -1, 1, 1, 1)
running_var = self.latent_norm_out.running_var.view(1, -1, 1, 1, 1)
eps = self.latent_norm_out.eps
z = z * torch.sqrt(running_var + eps) + running_mean
elif isinstance(self.latent_norm_out, nn.BatchNorm3d):
raise NotImplementedError("BatchNorm2d not supported")
return z
def _encode(self, x: torch.FloatTensor) -> AutoencoderKLOutput:
h = self.encoder(x)
moments = self.quant_conv(h)
moments = self._normalize_latent_channels(moments)
return moments
def _decode(
self,
z: torch.FloatTensor,
target_shape=None,
timestep: Optional[torch.Tensor] = None,
) -> Union[DecoderOutput, torch.FloatTensor]:
z = self._unnormalize_latent_channels(z)
z = self.post_quant_conv(z)
if "timestep" in self.decoder_params:
dec = self.decoder(z, target_shape=target_shape, timestep=timestep)
else:
dec = self.decoder(z, target_shape=target_shape)
return dec
def decode(
self,
z: torch.FloatTensor,
return_dict: bool = True,
target_shape=None,
timestep: Optional[torch.Tensor] = None,
) -> Union[DecoderOutput, torch.FloatTensor]:
assert target_shape is not None, "target_shape must be provided for decoding"
if self.use_z_tiling and z.shape[2] > self.z_sample_size > 1:
reduction_factor = int(
self.encoder.patch_size_t
* 2
** (
len(self.encoder.down_blocks)
- 1
- math.sqrt(self.encoder.patch_size)
)
)
split_size = self.z_sample_size // reduction_factor
num_splits = z.shape[2] // split_size
# copy target shape, and divide frame dimension (=2) by the context size
target_shape_split = list(target_shape)
target_shape_split[2] = target_shape[2] // num_splits
decoded_tiles = [
(
self._hw_tiled_decode(z_tile, target_shape_split)
if self.use_hw_tiling
else self._decode(z_tile, target_shape=target_shape_split)
)
for z_tile in torch.tensor_split(z, num_splits, dim=2)
]
decoded = torch.cat(decoded_tiles, dim=2)
else:
decoded = (
self._hw_tiled_decode(z, target_shape)
if self.use_hw_tiling
else self._decode(z, target_shape=target_shape, timestep=timestep)
)
if not return_dict:
return (decoded,)
return DecoderOutput(sample=decoded)
def forward(
self,
sample: torch.FloatTensor,
sample_posterior: bool = False,
return_dict: bool = True,
generator: Optional[torch.Generator] = None,
) -> Union[DecoderOutput, torch.FloatTensor]:
r"""
Args:
sample (`torch.FloatTensor`): Input sample.
sample_posterior (`bool`, *optional*, defaults to `False`):
Whether to sample from the posterior.
return_dict (`bool`, *optional*, defaults to `True`):
Whether to return a [`DecoderOutput`] instead of a plain tuple.
generator (`torch.Generator`, *optional*):
Generator used to sample from the posterior.
"""
x = sample
posterior = self.encode(x).latent_dist
if sample_posterior:
z = posterior.sample(generator=generator)
else:
z = posterior.mode()
dec = self.decode(z, target_shape=sample.shape).sample
if not return_dict:
return (dec,)
return DecoderOutput(sample=dec)
================================================
FILE: ltx_video/models/autoencoders/vae_encode.py
================================================
from typing import Tuple
import torch
from diffusers import AutoencoderKL
from einops import rearrange
from torch import Tensor
from ltx_video.models.autoencoders.causal_video_autoencoder import (
CausalVideoAutoencoder,
)
from ltx_video.models.autoencoders.video_autoencoder import (
Downsample3D,
VideoAutoencoder,
)
try:
import torch_xla.core.xla_model as xm
except ImportError:
xm = None
def vae_encode(
media_items: Tensor,
vae: AutoencoderKL,
split_size: int = 1,
vae_per_channel_normalize=False,
) -> Tensor:
"""
Encodes media items (images or videos) into latent representations using a specified VAE model.
The function supports processing batches of images or video frames and can handle the processing
in smaller sub-batches if needed.
Args:
media_items (Tensor): A torch Tensor containing the media items to encode. The expected
shape is (batch_size, channels, height, width) for images or (batch_size, channels,
frames, height, width) for videos.
vae (AutoencoderKL): An instance of the `AutoencoderKL` class from the `diffusers` library,
pre-configured and loaded with the appropriate model weights.
split_size (int, optional): The number of sub-batches to split the input batch into for encoding.
If set to more than 1, the input media items are processed in smaller batches according to
this value. Defaults to 1, which processes all items in a single batch.
Returns:
Tensor: A torch Tensor of the encoded latent representations. The shape of the tensor is adjusted
to match the input shape, scaled by the model's configuration.
Examples:
>>> import torch
>>> from diffusers import AutoencoderKL
>>> vae = AutoencoderKL.from_pretrained('your-model-name')
>>> images = torch.rand(10, 3, 8 256, 256) # Example tensor with 10 videos of 8 frames.
>>> latents = vae_encode(images, vae)
>>> print(latents.shape) # Output shape will depend on the model's latent configuration.
Note:
In case of a video, the function encodes the media item frame-by frame.
"""
is_video_shaped = media_items.dim() == 5
batch_size, channels = media_items.shape[0:2]
if channels != 3:
raise ValueError(f"Expects tensors with 3 channels, got {channels}.")
if is_video_shaped and not isinstance(
vae, (VideoAutoencoder, CausalVideoAutoencoder)
):
media_items = rearrange(media_items, "b c n h w -> (b n) c h w")
if split_size > 1:
if len(media_items) % split_size != 0:
raise ValueError(
"Error: The batch size must be divisible by 'train.vae_bs_split"
)
encode_bs = len(media_items) // split_size
# latents = [vae.encode(image_batch).latent_dist.sample() for image_batch in media_items.split(encode_bs)]
latents = []
if media_items.device.type == "xla":
xm.mark_step()
for image_batch in media_items.split(encode_bs):
latents.append(vae.encode(image_batch).latent_dist.sample())
if media_items.device.type == "xla":
xm.mark_step()
latents = torch.cat(latents, dim=0)
else:
latents = vae.encode(media_items).latent_dist.sample()
latents = normalize_latents(latents, vae, vae_per_channel_normalize)
if is_video_shaped and not isinstance(
vae, (VideoAutoencoder, CausalVideoAutoencoder)
):
latents = rearrange(latents, "(b n) c h w -> b c n h w", b=batch_size)
return latents
def vae_decode(
latents: Tensor,
vae: AutoencoderKL,
is_video: bool = True,
split_size: int = 1,
vae_per_channel_normalize=False,
timestep=None,
) -> Tensor:
is_video_shaped = latents.dim() == 5
batch_size = latents.shape[0]
if is_video_shaped and not isinstance(
vae, (VideoAutoencoder, CausalVideoAutoencoder)
):
latents = rearrange(latents, "b c n h w -> (b n) c h w")
if split_size > 1:
if len(latents) % split_size != 0:
raise ValueError(
"Error: The batch size must be divisible by 'train.vae_bs_split"
)
encode_bs = len(latents) // split_size
image_batch = [
_run_decoder(
latent_batch, vae, is_video, vae_per_channel_normalize, timestep
)
for latent_batch in latents.split(encode_bs)
]
images = torch.cat(image_batch, dim=0)
else:
images = _run_decoder(
latents, vae, is_video, vae_per_channel_normalize, timestep
)
if is_video_shaped and not isinstance(
vae, (VideoAutoencoder, CausalVideoAutoencoder)
):
images = rearrange(images, "(b n) c h w -> b c n h w", b=batch_size)
return images
def _run_decoder(
latents: Tensor,
vae: AutoencoderKL,
is_video: bool,
vae_per_channel_normalize=False,
timestep=None,
) -> Tensor:
if isinstance(vae, (VideoAutoencoder, CausalVideoAutoencoder)):
*_, fl, hl, wl = latents.shape
temporal_scale, spatial_scale, _ = get_vae_size_scale_factor(vae)
latents = latents.to(vae.dtype)
vae_decode_kwargs = {}
if timestep is not None:
vae_decode_kwargs["timestep"] = timestep
image = vae.decode(
un_normalize_latents(latents, vae, vae_per_channel_normalize),
return_dict=False,
target_shape=(
1,
3,
fl * temporal_scale if is_video else 1,
hl * spatial_scale,
wl * spatial_scale,
),
**vae_decode_kwargs,
)[0]
else:
image = vae.decode(
un_normalize_latents(latents, vae, vae_per_channel_normalize),
return_dict=False,
)[0]
return image
def get_vae_size_scale_factor(vae: AutoencoderKL) -> float:
if isinstance(vae, CausalVideoAutoencoder):
spatial = vae.spatial_downscale_factor
temporal = vae.temporal_downscale_factor
else:
down_blocks = len(
[
block
for block in vae.encoder.down_blocks
if isinstance(block.downsample, Downsample3D)
]
)
spatial = vae.config.patch_size * 2**down_blocks
temporal = (
vae.config.patch_size_t * 2**down_blocks
if isinstance(vae, VideoAutoencoder)
else 1
)
return (temporal, spatial, spatial)
def latent_to_pixel_coords(
latent_coords: Tensor, vae: AutoencoderKL, causal_fix: bool = False
) -> Tensor:
"""
Converts latent coordinates to pixel coordinates by scaling them according to the VAE's
configuration.
Args:
latent_coords (Tensor): A tensor of shape [batch_size, 3, num_latents]
containing the latent corner coordinates of each token.
vae (AutoencoderKL): The VAE model
causal_fix (bool): Whether to take into account the different temporal scale
of the first frame. Default = False for backwards compatibility.
Returns:
Tensor: A tensor of pixel coordinates corresponding to the input latent coordinates.
"""
scale_factors = get_vae_size_scale_factor(vae)
causal_fix = isinstance(vae, CausalVideoAutoencoder) and causal_fix
pixel_coords = latent_to_pixel_coords_from_factors(
latent_coords, scale_factors, causal_fix
)
return pixel_coords
def latent_to_pixel_coords_from_factors(
latent_coords: Tensor, scale_factors: Tuple, causal_fix: bool = False
) -> Tensor:
pixel_coords = (
latent_coords
* torch.tensor(scale_factors, device=latent_coords.device)[None, :, None]
)
if causal_fix:
# Fix temporal scale for first frame to 1 due to causality
pixel_coords[:, 0] = (pixel_coords[:, 0] + 1 - scale_factors[0]).clamp(min=0)
return pixel_coords
def normalize_latents(
latents: Tensor, vae: AutoencoderKL, vae_per_channel_normalize: bool = False
) -> Tensor:
return (
(latents - vae.mean_of_means.to(latents.dtype).view(1, -1, 1, 1, 1))
/ vae.std_of_means.to(latents.dtype).view(1, -1, 1, 1, 1)
if vae_per_channel_normalize
else latents * vae.config.scaling_factor
)
def un_normalize_latents(
latents: Tensor, vae: AutoencoderKL, vae_per_channel_normalize: bool = False
) -> Tensor:
return (
latents * vae.std_of_means.to(latents.dtype).view(1, -1, 1, 1, 1)
+ vae.mean_of_means.to(latents.dtype).view(1, -1, 1, 1, 1)
if vae_per_channel_normalize
else latents / vae.config.scaling_factor
)
================================================
FILE: ltx_video/models/autoencoders/video_autoencoder.py
================================================
import json
import os
from functools import partial
from types import SimpleNamespace
from typing import Any, Mapping, Optional, Tuple, Union
import torch
from einops import rearrange
from torch import nn
from torch.nn import functional
from diffusers.utils import logging
from ltx_video.utils.torch_utils import Identity
from ltx_video.models.autoencoders.conv_nd_factory import make_conv_nd, make_linear_nd
from ltx_video.models.autoencoders.pixel_norm import PixelNorm
from ltx_video.models.autoencoders.vae import AutoencoderKLWrapper
logger = logging.get_logger(__name__)
class VideoAutoencoder(AutoencoderKLWrapper):
@classmethod
def from_pretrained(
cls,
pretrained_model_name_or_path: Optional[Union[str, os.PathLike]],
*args,
**kwargs,
):
config_local_path = pretrained_model_name_or_path / "config.json"
config = cls.load_config(config_local_path, **kwargs)
video_vae = cls.from_config(config)
video_vae.to(kwargs["torch_dtype"])
model_local_path = pretrained_model_name_or_path / "autoencoder.pth"
ckpt_state_dict = torch.load(model_local_path)
video_vae.load_state_dict(ckpt_state_dict)
statistics_local_path = (
pretrained_model_name_or_path / "per_channel_statistics.json"
)
if statistics_local_path.exists():
with open(statistics_local_path, "r") as file:
data = json.load(file)
transposed_data = list(zip(*data["data"]))
data_dict = {
col: torch.tensor(vals)
for col, vals in zip(data["columns"], transposed_data)
}
video_vae.register_buffer("std_of_means", data_dict["std-of-means"])
video_vae.register_buffer(
"mean_of_means",
data_dict.get(
"mean-of-means", torch.zeros_like(data_dict["std-of-means"])
),
)
return video_vae
@staticmethod
def from_config(config):
assert (
config["_class_name"] == "VideoAutoencoder"
), "config must have _class_name=VideoAutoencoder"
if isinstance(config["dims"], list):
config["dims"] = tuple(config["dims"])
assert config["dims"] in [2, 3, (2, 1)], "dims must be 2, 3 or (2, 1)"
double_z = config.get("double_z", True)
latent_log_var = config.get(
"latent_log_var", "per_channel" if double_z else "none"
)
use_quant_conv = config.get("use_quant_conv", True)
if use_quant_conv and latent_log_var == "uniform":
raise ValueError("uniform latent_log_var requires use_quant_conv=False")
encoder = Encoder(
dims=config["dims"],
in_channels=config.get("in_channels", 3),
out_channels=config["latent_channels"],
block_out_channels=config["block_out_channels"],
patch_size=config.get("patch_size", 1),
latent_log_var=latent_log_var,
norm_layer=config.get("norm_layer", "group_norm"),
patch_size_t=config.get("patch_size_t", config.get("patch_size", 1)),
add_channel_padding=config.get("add_channel_padding", False),
)
decoder = Decoder(
dims=config["dims"],
in_channels=config["latent_channels"],
out_channels=config.get("out_channels", 3),
block_out_channels=config["block_out_channels"],
patch_size=config.get("patch_size", 1),
norm_layer=config.get("norm_layer", "group_norm"),
patch_size_t=config.get("patch_size_t", config.get("patch_size", 1)),
add_channel_padding=config.get("add_channel_padding", False),
)
dims = config["dims"]
return VideoAutoencoder(
encoder=encoder,
decoder=decoder,
latent_channels=config["latent_channels"],
dims=dims,
use_quant_conv=use_quant_conv,
)
@property
def config(self):
return SimpleNamespace(
_class_name="VideoAutoencoder",
dims=self.dims,
in_channels=self.encoder.conv_in.in_channels
// (self.encoder.patch_size_t * self.encoder.patch_size**2),
out_channels=self.decoder.conv_out.out_channels
// (self.decoder.patch_size_t * self.decoder.patch_size**2),
latent_channels=self.decoder.conv_in.in_channels,
block_out_channels=[
self.encoder.down_blocks[i].res_blocks[-1].conv1.out_channels
for i in range(len(self.encoder.down_blocks))
],
scaling_factor=1.0,
norm_layer=self.encoder.norm_layer,
patch_size=self.encoder.patch_size,
latent_log_var=self.encoder.latent_log_var,
use_quant_conv=self.use_quant_conv,
patch_size_t=self.encoder.patch_size_t,
add_channel_padding=self.encoder.add_channel_padding,
)
@property
def is_video_supported(self):
"""
Check if the model supports video inputs of shape (B, C, F, H, W). Otherwise, the model only supports 2D images.
"""
return self.dims != 2
@property
def downscale_factor(self):
return self.encoder.downsample_factor
def to_json_string(self) -> str:
import json
return json.dumps(self.config.__dict__)
def load_state_dict(self, state_dict: Mapping[str, Any], strict: bool = True):
model_keys = set(name for name, _ in self.named_parameters())
key_mapping = {
".resnets.": ".res_blocks.",
"downsamplers.0": "downsample",
"upsamplers.0": "upsample",
}
converted_state_dict = {}
for key, value in state_dict.items():
for k, v in key_mapping.items():
key = key.replace(k, v)
if "norm" in key and key not in model_keys:
logger.info(
f"Removing key {key} from state_dict as it is not present in the model"
)
continue
converted_state_dict[key] = value
super().load_state_dict(converted_state_dict, strict=strict)
def last_layer(self):
if hasattr(self.decoder, "conv_out"):
if isinstance(self.decoder.conv_out, nn.Sequential):
last_layer = self.decoder.conv_out[-1]
else:
last_layer = self.decoder.conv_out
else:
last_layer = self.decoder.layers[-1]
return last_layer
class Encoder(nn.Module):
r"""
The `Encoder` layer of a variational autoencoder that encodes its input into a latent representation.
Args:
in_channels (`int`, *optional*, defaults to 3):
The number of input channels.
out_channels (`int`, *optional*, defaults to 3):
The number of output channels.
block_out_channels (`Tuple[int, ...]`, *optional*, defaults to `(64,)`):
The number of output channels for each block.
layers_per_block (`int`, *optional*, defaults to 2):
The number of layers per block.
norm_num_groups (`int`, *optional*, defaults to 32):
The number of groups for normalization.
patch_size (`int`, *optional*, defaults to 1):
The patch size to use. Should be a power of 2.
norm_layer (`str`, *optional*, defaults to `group_norm`):
The normalization layer to use. Can be either `group_norm` or `pixel_norm`.
latent_log_var (`str`, *optional*, defaults to `per_channel`):
The number of channels for the log variance. Can be either `per_channel`, `uniform`, or `none`.
"""
def __init__(
self,
dims: Union[int, Tuple[int, int]] = 3,
in_channels: int = 3,
out_channels: int = 3,
block_out_channels: Tuple[int, ...] = (64,),
layers_per_block: int = 2,
norm_num_groups: int = 32,
patch_size: Union[int, Tuple[int]] = 1,
norm_layer: str = "group_norm", # group_norm, pixel_norm
latent_log_var: str = "per_channel",
patch_size_t: Optional[int] = None,
add_channel_padding: Optional[bool] = False,
):
super().__init__()
self.patch_size = patch_size
self.patch_size_t = patch_size_t if patch_size_t is not None else patch_size
self.add_channel_padding = add_channel_padding
self.layers_per_block = layers_per_block
self.norm_layer = norm_layer
self.latent_channels = out_channels
self.latent_log_var = latent_log_var
if add_channel_padding:
in_channels = in_channels * self.patch_size**3
else:
in_channels = in_channels * self.patch_size_t * self.patch_size**2
self.in_channels = in_channels
output_channel = block_out_channels[0]
self.conv_in = make_conv_nd(
dims=dims,
in_channels=in_channels,
out_channels=output_channel,
kernel_size=3,
stride=1,
padding=1,
)
self.down_blocks = nn.ModuleList([])
for i in range(len(block_out_channels)):
input_channel = output_channel
output_channel = block_out_channels[i]
is_final_block = i == len(block_out_channels) - 1
down_block = DownEncoderBlock3D(
dims=dims,
in_channels=input_channel,
out_channels=output_channel,
num_layers=self.layers_per_block,
add_downsample=not is_final_block and 2**i >= patch_size,
resnet_eps=1e-6,
downsample_padding=0,
resnet_groups=norm_num_groups,
norm_layer=norm_layer,
)
self.down_blocks.append(down_block)
self.mid_block = UNetMidBlock3D(
dims=dims,
in_channels=block_out_channels[-1],
num_layers=self.layers_per_block,
resnet_eps=1e-6,
resnet_groups=norm_num_groups,
norm_layer=norm_layer,
)
# out
if norm_layer == "group_norm":
self.conv_norm_out = nn.GroupNorm(
num_channels=block_out_channels[-1],
num_groups=norm_num_groups,
eps=1e-6,
)
elif norm_layer == "pixel_norm":
self.conv_norm_out = PixelNorm()
self.conv_act = nn.SiLU()
conv_out_channels = out_channels
if latent_log_var == "per_channel":
conv_out_channels *= 2
elif latent_log_var == "uniform":
conv_out_channels += 1
elif latent_log_var != "none":
raise ValueError(f"Invalid latent_log_var: {latent_log_var}")
self.conv_out = make_conv_nd(
dims, block_out_channels[-1], conv_out_channels, 3, padding=1
)
self.gradient_checkpointing = False
@property
def downscale_factor(self):
return (
2
** len(
[
block
for block in self.down_blocks
if isinstance(block.downsample, Downsample3D)
]
)
* self.patch_size
)
def forward(
self, sample: torch.FloatTensor, return_features=False
) -> torch.FloatTensor:
r"""The forward method of the `Encoder` class."""
downsample_in_time = sample.shape[2] != 1
# patchify
patch_size_t = self.patch_size_t if downsample_in_time else 1
sample = patchify(
sample,
patch_size_hw=self.patch_size,
patch_size_t=patch_size_t,
add_channel_padding=self.add_channel_padding,
)
sample = self.conv_in(sample)
checkpoint_fn = (
partial(torch.utils.checkpoint.checkpoint, use_reentrant=False)
if self.gradient_checkpointing and self.training
else lambda x: x
)
if return_features:
features = []
for down_block in self.down_blocks:
sample = checkpoint_fn(down_block)(
sample, downsample_in_time=downsample_in_time
)
if return_features:
features.append(sample)
sample = checkpoint_fn(self.mid_block)(sample)
# post-process
sample = self.conv_norm_out(sample)
sample = self.conv_act(sample)
sample = self.conv_out(sample)
if self.latent_log_var == "uniform":
last_channel = sample[:, -1:, ...]
num_dims = sample.dim()
if num_dims == 4:
# For shape (B, C, H, W)
repeated_last_channel = last_channel.repeat(
1, sample.shape[1] - 2, 1, 1
)
sample = torch.cat([sample, repeated_last_channel], dim=1)
elif num_dims == 5:
# For shape (B, C, F, H, W)
repeated_last_channel = last_channel.repeat(
1, sample.shape[1] - 2, 1, 1, 1
)
sample = torch.cat([sample, repeated_last_channel], dim=1)
else:
raise ValueError(f"Invalid input shape: {sample.shape}")
if return_features:
features.append(sample[:, : self.latent_channels, ...])
return sample, features
return sample
class Decoder(nn.Module):
r"""
The `Decoder` layer of a variational autoencoder that decodes its latent representation into an output sample.
Args:
in_channels (`int`, *optional*, defaults to 3):
The number of input channels.
out_channels (`int`, *optional*, defaults to 3):
The number of output channels.
block_out_channels (`Tuple[int, ...]`, *optional*, defaults to `(64,)`):
The number of output channels for each block.
layers_per_block (`int`, *optional*, defaults to 2):
The number of layers per block.
norm_num_groups (`int`, *optional*, defaults to 32):
The number of groups for normalization.
patch_size (`int`, *optional*, defaults to 1):
The patch size to use. Should be a power of 2.
norm_layer (`str`, *optional*, defaults to `group_norm`):
The normalization layer to use. Can be either `group_norm` or `pixel_norm`.
"""
def __init__(
self,
dims,
in_channels: int = 3,
out_channels: int = 3,
block_out_channels: Tuple[int, ...] = (64,),
layers_per_block: int = 2,
norm_num_groups: int = 32,
patch_size: int = 1,
norm_layer: str = "group_norm",
patch_size_t: Optional[int] = None,
add_channel_padding: Optional[bool] = False,
):
super().__init__()
self.patch_size = patch_size
self.patch_size_t = patch_size_t if patch_size_t is not None else patch_size
self.add_channel_padding = add_channel_padding
self.layers_per_block = layers_per_block
if add_channel_padding:
out_channels = out_channels * self.patch_size**3
else:
out_channels = out_channels * self.patch_size_t * self.patch_size**2
self.out_channels = out_channels
self.conv_in = make_conv_nd(
dims,
in_channels,
block_out_channels[-1],
kernel_size=3,
stride=1,
padding=1,
)
self.mid_block = None
self.up_blocks = nn.ModuleList([])
self.mid_block = UNetMidBlock3D(
dims=dims,
in_channels=block_out_channels[-1],
num_layers=self.layers_per_block,
resnet_eps=1e-6,
resnet_groups=norm_num_groups,
norm_layer=norm_layer,
)
reversed_block_out_channels = list(reversed(block_out_channels))
output_channel = reversed_block_out_channels[0]
for i in range(len(reversed_block_out_channels)):
prev_output_channel = output_channel
output_channel = reversed_block_out_channels[i]
is_final_block = i == len(block_out_channels) - 1
up_block = UpDecoderBlock3D(
dims=dims,
num_layers=self.layers_per_block + 1,
in_channels=prev_output_channel,
out_channels=output_channel,
add_upsample=not is_final_block
and 2 ** (len(block_out_channels) - i - 1) > patch_size,
resnet_eps=1e-6,
resnet_groups=norm_num_groups,
norm_layer=norm_layer,
)
self.up_blocks.append(up_block)
if norm_layer == "group_norm":
self.conv_norm_out = nn.GroupNorm(
num_channels=block_out_channels[0], num_groups=norm_num_groups, eps=1e-6
)
elif norm_layer == "pixel_norm":
self.conv_norm_out = PixelNorm()
self.conv_act = nn.SiLU()
self.conv_out = make_conv_nd(
dims, block_out_channels[0], out_channels, 3, padding=1
)
self.gradient_checkpointing = False
def forward(self, sample: torch.FloatTensor, target_shape) -> torch.FloatTensor:
r"""The forward method of the `Decoder` class."""
assert target_shape is not None, "target_shape must be provided"
upsample_in_time = sample.shape[2] < target_shape[2]
sample = self.conv_in(sample)
upscale_dtype = next(iter(self.up_blocks.parameters())).dtype
checkpoint_fn = (
partial(torch.utils.checkpoint.checkpoint, use_reentrant=False)
if self.gradient_checkpointing and self.training
else lambda x: x
)
sample = checkpoint_fn(self.mid_block)(sample)
sample = sample.to(upscale_dtype)
for up_block in self.up_blocks:
sample = checkpoint_fn(up_block)(sample, upsample_in_time=upsample_in_time)
# post-process
sample = self.conv_norm_out(sample)
sample = self.conv_act(sample)
sample = self.conv_out(sample)
# un-patchify
patch_size_t = self.patch_size_t if upsample_in_time else 1
sample = unpatchify(
sample,
patch_size_hw=self.patch_size,
patch_size_t=patch_size_t,
add_channel_padding=self.add_channel_padding,
)
return sample
class DownEncoderBlock3D(nn.Module):
def __init__(
self,
dims: Union[int, Tuple[int, int]],
in_channels: int,
out_channels: int,
dropout: float = 0.0,
num_layers: int = 1,
resnet_eps: float = 1e-6,
resnet_groups: int = 32,
add_downsample: bool = True,
downsample_padding: int = 1,
norm_layer: str = "group_norm",
):
super().__init__()
res_blocks = []
for i in range(num_layers):
in_channels = in_channels if i == 0 else out_channels
res_blocks.append(
ResnetBlock3D(
dims=dims,
in_channels=in_channels,
out_channels=out_channels,
eps=resnet_eps,
groups=resnet_groups,
dropout=dropout,
norm_layer=norm_layer,
)
)
self.res_blocks = nn.ModuleList(res_blocks)
if add_downsample:
self.downsample = Downsample3D(
dims,
out_channels,
out_channels=out_channels,
padding=downsample_padding,
)
else:
self.downsample = Identity()
def forward(
self, hidden_states: torch.FloatTensor, downsample_in_time
) -> torch.FloatTensor:
for resnet in self.res_blocks:
hidden_states = resnet(hidden_states)
hidden_states = self.downsample(
hidden_states, downsample_in_time=downsample_in_time
)
return hidden_states
class UNetMidBlock3D(nn.Module):
"""
A 3D UNet mid-block [`UNetMidBlock3D`] with multiple residual blocks.
Args:
in_channels (`int`): The number of input channels.
dropout (`float`, *optional*, defaults to 0.0): The dropout rate.
num_layers (`int`, *optional*, defaults to 1): The number of residual blocks.
resnet_eps (`float`, *optional*, 1e-6 ): The epsilon value for the resnet blocks.
resnet_groups (`int`, *optional*, defaults to 32):
The number of groups to use in the group normalization layers of the resnet blocks.
Returns:
`torch.FloatTensor`: The output of the last residual block, which is a tensor of shape `(batch_size,
in_channels, height, width)`.
"""
def __init__(
self,
dims: Union[int, Tuple[int, int]],
in_channels: int,
dropout: float = 0.0,
num_layers: int = 1,
resnet_eps: float = 1e-6,
resnet_groups: int = 32,
norm_layer: str = "group_norm",
):
super().__init__()
resnet_groups = (
resnet_groups if resnet_groups is not None else min(in_channels // 4, 32)
)
self.res_blocks = nn.ModuleList(
[
ResnetBlock3D(
dims=dims,
in_channels=in_channels,
out_channels=in_channels,
eps=resnet_eps,
groups=resnet_groups,
dropout=dropout,
norm_layer=norm_layer,
)
for _ in range(num_layers)
]
)
def forward(self, hidden_states: torch.FloatTensor) -> torch.FloatTensor:
for resnet in self.res_blocks:
hidden_states = resnet(hidden_states)
return hidden_states
class UpDecoderBlock3D(nn.Module):
def __init__(
self,
dims: Union[int, Tuple[int, int]],
in_channels: int,
out_channels: int,
resolution_idx: Optional[int] = None,
dropout: float = 0.0,
num_layers: int = 1,
resnet_eps: float = 1e-6,
resnet_groups: int = 32,
add_upsample: bool = True,
norm_layer: str = "group_norm",
):
super().__init__()
res_blocks = []
for i in range(num_layers):
input_channels = in_channels if i == 0 else out_channels
res_blocks.append(
ResnetBlock3D(
dims=dims,
in_channels=input_channels,
out_channels=out_channels,
eps=resnet_eps,
groups=resnet_groups,
dropout=dropout,
norm_layer=norm_layer,
)
)
self.res_blocks = nn.ModuleList(res_blocks)
if add_upsample:
self.upsample = Upsample3D(
dims=dims, channels=out_channels, out_channels=out_channels
)
else:
self.upsample = Identity()
self.resolution_idx = resolution_idx
def forward(
self, hidden_states: torch.FloatTensor, upsample_in_time=True
) -> torch.FloatTensor:
for resnet in self.res_blocks:
hidden_states = resnet(hidden_states)
hidden_states = self.upsample(hidden_states, upsample_in_time=upsample_in_time)
return hidden_states
class ResnetBlock3D(nn.Module):
r"""
A Resnet block.
Parameters:
in_channels (`int`): The number of channels in the input.
out_channels (`int`, *optional*, default to be `None`):
The number of output channels for the first conv layer. If None, same as `in_channels`.
dropout (`float`, *optional*, defaults to `0.0`): The dropout probability to use.
groups (`int`, *optional*, default to `32`): The number of groups to use for the first normalization layer.
eps (`float`, *optional*, defaults to `1e-6`): The epsilon to use for the normalization.
"""
def __init__(
self,
dims: Union[int, Tuple[int, int]],
in_channels: int,
out_channels: Optional[int] = None,
conv_shortcut: bool = False,
dropout: float = 0.0,
groups: int = 32,
eps: float = 1e-6,
norm_layer: str = "group_norm",
):
super().__init__()
self.in_channels = in_channels
out_channels = in_channels if out_channels is None else out_channels
self.out_channels = out_channels
self.use_conv_shortcut = conv_shortcut
if norm_layer == "group_norm":
self.norm1 = torch.nn.GroupNorm(
num_groups=groups, num_channels=in_channels, eps=eps, affine=True
)
elif norm_layer == "pixel_norm":
self.norm1 = PixelNorm()
self.non_linearity = nn.SiLU()
self.conv1 = make_conv_nd(
dims, in_channels, out_channels, kernel_size=3, stride=1, padding=1
)
if norm_layer == "group_norm":
self.norm2 = torch.nn.GroupNorm(
num_groups=groups, num_channels=out_channels, eps=eps, affine=True
)
elif norm_layer == "pixel_norm":
self.norm2 = PixelNorm()
self.dropout = torch.nn.Dropout(dropout)
self.conv2 = make_conv_nd(
dims, out_channels, out_channels, kernel_size=3, stride=1, padding=1
)
self.conv_shortcut = (
make_linear_nd(
dims=dims, in_channels=in_channels, out_channels=out_channels
)
if in_channels != out_channels
else nn.Identity()
)
def forward(
self,
input_tensor: torch.FloatTensor,
) -> torch.FloatTensor:
hidden_states = input_tensor
hidden_states = self.norm1(hidden_states)
hidden_states = self.non_linearity(hidden_states)
hidden_states = self.conv1(hidden_states)
hidden_states = self.norm2(hidden_states)
hidden_states = self.non_linearity(hidden_states)
hidden_states = self.dropout(hidden_states)
hidden_states = self.conv2(hidden_states)
input_tensor = self.conv_shortcut(input_tensor)
output_tensor = input_tensor + hidden_states
return output_tensor
class Downsample3D(nn.Module):
def __init__(
self,
dims,
in_channels: int,
out_channels: int,
kernel_size: int = 3,
padding: int = 1,
):
super().__init__()
stride: int = 2
self.padding = padding
self.in_channels = in_channels
self.dims = dims
self.conv = make_conv_nd(
dims=dims,
in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
)
def forward(self, x, downsample_in_time=True):
conv = self.conv
if self.padding == 0:
if self.dims == 2:
padding = (0, 1, 0, 1)
else:
padding = (0, 1, 0, 1, 0, 1 if downsample_in_time else 0)
x = functional.pad(x, padding, mode="constant", value=0)
if self.dims == (2, 1) and not downsample_in_time:
return conv(x, skip_time_conv=True)
return conv(x)
class Upsample3D(nn.Module):
"""
An upsampling layer for 3D tensors of shape (B, C, D, H, W).
:param channels: channels in the inputs and outputs.
"""
def __init__(self, dims, channels, out_channels=None):
super().__init__()
self.dims = dims
self.channels = channels
self.out_channels = out_channels or channels
self.conv = make_conv_nd(
dims, channels, out_channels, kernel_size=3, padding=1, bias=True
)
def forward(self, x, upsample_in_time):
if self.dims == 2:
x = functional.interpolate(
x, (x.shape[2] * 2, x.shape[3] * 2), mode="nearest"
)
else:
time_scale_factor = 2 if upsample_in_time else 1
# print("before:", x.shape)
b, c, d, h, w = x.shape
x = rearrange(x, "b c d h w -> (b d) c h w")
# height and width interpolate
x = functional.interpolate(
x, (x.shape[2] * 2, x.shape[3] * 2), mode="nearest"
)
_, _, h, w = x.shape
if not upsample_in_time and self.dims == (2, 1):
x = rearrange(x, "(b d) c h w -> b c d h w ", b=b, h=h, w=w)
return self.conv(x, skip_time_conv=True)
# Second ** upsampling ** which is essentially treated as a 1D convolution across the 'd' dimension
x = rearrange(x, "(b d) c h w -> (b h w) c 1 d", b=b)
# (b h w) c 1 d
new_d = x.shape[-1] * time_scale_factor
x = functional.interpolate(x, (1, new_d), mode="nearest")
# (b h w) c 1 new_d
x = rearrange(
x, "(b h w) c 1 new_d -> b c new_d h w", b=b, h=h, w=w, new_d=new_d
)
# b c d h w
# x = functional.interpolate(
# x, (x.shape[2] * time_scale_factor, x.shape[3] * 2, x.shape[4] * 2), mode="nearest"
# )
# print("after:", x.shape)
return self.conv(x)
def patchify(x, patch_size_hw, patch_size_t=1, add_channel_padding=False):
if patch_size_hw == 1 and patch_size_t == 1:
return x
if x.dim() == 4:
x = rearrange(
x, "b c (h q) (w r) -> b (c r q) h w", q=patch_size_hw, r=patch_size_hw
)
elif x.dim() == 5:
x = rearrange(
x,
"b c (f p) (h q) (w r) -> b (c p r q) f h w",
p=patch_size_t,
q=patch_size_hw,
r=patch_size_hw,
)
else:
raise ValueError(f"Invalid input shape: {x.shape}")
if (
(x.dim() == 5)
and (patch_size_hw > patch_size_t)
and (patch_size_t > 1 or add_channel_padding)
):
channels_to_pad = x.shape[1] * (patch_size_hw // patch_size_t) - x.shape[1]
padding_zeros = torch.zeros(
x.shape[0],
channels_to_pad,
x.shape[2],
x.shape[3],
x.shape[4],
device=x.device,
dtype=x.dtype,
)
x = torch.cat([padding_zeros, x], dim=1)
return x
def unpatchify(x, patch_size_hw, patch_size_t=1, add_channel_padding=False):
if patch_size_hw == 1 and patch_size_t == 1:
return x
if (
(x.dim() == 5)
and (patch_size_hw > patch_size_t)
and (patch_size_t > 1 or add_channel_padding)
):
channels_to_keep = int(x.shape[1] * (patch_size_t / patch_size_hw))
x = x[:, :channels_to_keep, :, :, :]
if x.dim() == 4:
x = rearrange(
x, "b (c r q) h w -> b c (h q) (w r)", q=patch_size_hw, r=patch_size_hw
)
elif x.dim() == 5:
x = rearrange(
x,
"b (c p r q) f h w -> b c (f p) (h q) (w r)",
p=patch_size_t,
q=patch_size_hw,
r=patch_size_hw,
)
return x
def create_video_autoencoder_config(
latent_channels: int = 4,
):
config = {
"_class_name": "VideoAutoencoder",
"dims": (
2,
1,
), # 2 for Conv2, 3 for Conv3d, (2, 1) for Conv2d followed by Conv1d
"in_channels": 3, # Number of input color channels (e.g., RGB)
"out_channels": 3, # Number of output color channels
"latent_channels": latent_channels, # Number of channels in the latent space representation
"block_out_channels": [
128,
256,
512,
512,
], # Number of output channels of each encoder / decoder inner block
"patch_size": 1,
}
return config
def create_video_autoencoder_pathify4x4x4_config(
latent_channels: int = 4,
):
config = {
"_class_name": "VideoAutoencoder",
"dims": (
2,
1,
), # 2 for Conv2, 3 for Conv3d, (2, 1) for Conv2d followed by Conv1d
"in_channels": 3, # Number of input color channels (e.g., RGB)
"out_channels": 3, # Number of output color channels
"latent_channels": latent_channels, # Number of channels in the latent space representation
"block_out_channels": [512]
* 4, # Number of output channels of each encoder / decoder inner block
"patch_size": 4,
"latent_log_var": "uniform",
}
return config
def create_video_autoencoder_pathify4x4_config(
latent_channels: int = 4,
):
config = {
"_class_name": "VideoAutoencoder",
"dims": 2, # 2 for Conv2, 3 for Conv3d, (2, 1) for Conv2d followed by Conv1d
"in_channels": 3, # Number of input color channels (e.g., RGB)
"out_channels": 3, # Number of output color channels
"latent_channels": latent_channels, # Number of channels in the latent space representation
"block_out_channels": [512]
* 4, # Number of output channels of each encoder / decoder inner block
"patch_size": 4,
"norm_layer": "pixel_norm",
}
return config
def test_vae_patchify_unpatchify():
import torch
x = torch.randn(2, 3, 8, 64, 64)
x_patched = patchify(x, patch_size_hw=4, patch_size_t=4)
x_unpatched = unpatchify(x_patched, patch_size_hw=4, patch_size_t=4)
assert torch.allclose(x, x_unpatched)
def demo_video_autoencoder_forward_backward():
# Configuration for the VideoAutoencoder
config = create_video_autoencoder_pathify4x4x4_config()
# Instantiate the VideoAutoencoder with the specified configuration
video_autoencoder = VideoAutoencoder.from_config(config)
print(video_autoencoder)
# Print the total number of parameters in the video autoencoder
total_params = sum(p.numel() for p in video_autoencoder.parameters())
print(f"Total number of parameters in VideoAutoencoder: {total_params:,}")
# Create a mock input tensor simulating a batch of videos
# Shape: (batch_size, channels, depth, height, width)
# E.g., 4 videos, each with 3 color channels, 16 frames, and 64x64 pixels per frame
input_videos = torch.randn(2, 3, 8, 64, 64)
# Forward pass: encode and decode the input videos
latent = video_autoencoder.encode(input_videos).latent_dist.mode()
print(f"input shape={input_videos.shape}")
print(f"latent shape={latent.shape}")
reconstructed_videos = video_autoencoder.decode(
latent, target_shape=input_videos.shape
).sample
print(f"reconstructed shape={reconstructed_videos.shape}")
# Calculate the loss (e.g., mean squared error)
loss = torch.nn.functional.mse_loss(input_videos, reconstructed_videos)
# Perform backward pass
loss.backward()
print(f"Demo completed with loss: {loss.item()}")
# Ensure to call the demo function to execute the forward and backward pass
if __name__ == "__main__":
demo_video_autoencoder_forward_backward()
================================================
FILE: ltx_video/models/transformers/__init__.py
================================================
================================================
FILE: ltx_video/models/transformers/attention.py
================================================
import inspect
from importlib import import_module
from typing import Any, Dict, Optional, Tuple
import torch
import torch.nn.functional as F
from diffusers.models.activations import GEGLU, GELU, ApproximateGELU
from diffusers.models.attention import _chunked_feed_forward
from diffusers.models.attention_processor import (
LoRAAttnAddedKVProcessor,
LoRAAttnProcessor,
LoRAAttnProcessor2_0,
LoRAXFormersAttnProcessor,
SpatialNorm,
)
from diffusers.models.lora import LoRACompatibleLinear
from diffusers.models.normalization import RMSNorm
from diffusers.utils import deprecate, logging
from diffusers.utils.torch_utils import maybe_allow_in_graph
from einops import rearrange
from torch import nn
from ltx_video.utils.skip_layer_strategy import SkipLayerStrategy
try:
from torch_xla.experimental.custom_kernel import flash_attention
except ImportError:
# workaround for automatic tests. Currently this function is manually patched
# to the torch_xla lib on setup of container
pass
# code adapted from https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention.py
logger = logging.get_logger(__name__)
@maybe_allow_in_graph
class BasicTransformerBlock(nn.Module):
r"""
A basic Transformer block.
Parameters:
dim (`int`): The number of channels in the input and output.
num_attention_heads (`int`): The number of heads to use for multi-head attention.
attention_head_dim (`int`): The number of channels in each head.
dropout (`float`, *optional*, defaults to 0.0): The dropout probability to use.
cross_attention_dim (`int`, *optional*): The size of the encoder_hidden_states vector for cross attention.
activation_fn (`str`, *optional*, defaults to `"geglu"`): Activation function to be used in feed-forward.
num_embeds_ada_norm (:
obj: `int`, *optional*): The number of diffusion steps used during training. See `Transformer2DModel`.
attention_bias (:
obj: `bool`, *optional*, defaults to `False`): Configure if the attentions should contain a bias parameter.
only_cross_attention (`bool`, *optional*):
Whether to use only cross-attention layers. In this case two cross attention layers are used.
double_self_attention (`bool`, *optional*):
Whether to use two self-attention layers. In this case no cross attention layers are used.
upcast_attention (`bool`, *optional*):
Whether to upcast the attention computation to float32. This is useful for mixed precision training.
norm_elementwise_affine (`bool`, *optional*, defaults to `True`):
Whether to use learnable elementwise affine parameters for normalization.
qk_norm (`str`, *optional*, defaults to None):
Set to 'layer_norm' or `rms_norm` to perform query and key normalization.
adaptive_norm (`str`, *optional*, defaults to `"single_scale_shift"`):
The type of adaptive norm to use. Can be `"single_scale_shift"`, `"single_scale"` or "none".
standardization_norm (`str`, *optional*, defaults to `"layer_norm"`):
The type of pre-normalization to use. Can be `"layer_norm"` or `"rms_norm"`.
final_dropout (`bool` *optional*, defaults to False):
Whether to apply a final dropout after the last feed-forward layer.
attention_type (`str`, *optional*, defaults to `"default"`):
The type of attention to use. Can be `"default"` or `"gated"` or `"gated-text-image"`.
positional_embeddings (`str`, *optional*, defaults to `None`):
The type of positional embeddings to apply to.
num_positional_embeddings (`int`, *optional*, defaults to `None`):
The maximum number of positional embeddings to apply.
"""
def __init__(
self,
dim: int,
num_attention_heads: int,
attention_head_dim: int,
dropout=0.0,
cross_attention_dim: Optional[int] = None,
activation_fn: str = "geglu",
num_embeds_ada_norm: Optional[int] = None, # pylint: disable=unused-argument
attention_bias: bool = False,
only_cross_attention: bool = False,
double_self_attention: bool = False,
upcast_attention: bool = False,
norm_elementwise_affine: bool = True,
adaptive_norm: str = "single_scale_shift", # 'single_scale_shift', 'single_scale' or 'none'
standardization_norm: str = "layer_norm", # 'layer_norm' or 'rms_norm'
norm_eps: float = 1e-5,
qk_norm: Optional[str] = None,
final_dropout: bool = False,
attention_type: str = "default", # pylint: disable=unused-argument
ff_inner_dim: Optional[int] = None,
ff_bias: bool = True,
attention_out_bias: bool = True,
use_tpu_flash_attention: bool = False,
use_rope: bool = False,
):
super().__init__()
self.only_cross_attention = only_cross_attention
self.use_tpu_flash_attention = use_tpu_flash_attention
self.adaptive_norm = adaptive_norm
assert standardization_norm in ["layer_norm", "rms_norm"]
assert adaptive_norm in ["single_scale_shift", "single_scale", "none"]
make_norm_layer = (
nn.LayerNorm if standardization_norm == "layer_norm" else RMSNorm
)
# Define 3 blocks. Each block has its own normalization layer.
# 1. Self-Attn
self.norm1 = make_norm_layer(
dim, elementwise_affine=norm_elementwise_affine, eps=norm_eps
)
self.attn1 = Attention(
query_dim=dim,
heads=num_attention_heads,
dim_head=attention_head_dim,
dropout=dropout,
bias=attention_bias,
cross_attention_dim=cross_attention_dim if only_cross_attention else None,
upcast_attention=upcast_attention,
out_bias=attention_out_bias,
use_tpu_flash_attention=use_tpu_flash_attention,
qk_norm=qk_norm,
use_rope=use_rope,
)
# 2. Cross-Attn
if cross_attention_dim is not None or double_self_attention:
self.attn2 = Attention(
query_dim=dim,
cross_attention_dim=(
cross_attention_dim if not double_self_attention else None
),
heads=num_attention_heads,
dim_head=attention_head_dim,
dropout=dropout,
bias=attention_bias,
upcast_attention=upcast_attention,
out_bias=attention_out_bias,
use_tpu_flash_attention=use_tpu_flash_attention,
qk_norm=qk_norm,
use_rope=use_rope,
) # is self-attn if encoder_hidden_states is none
if adaptive_norm == "none":
self.attn2_norm = make_norm_layer(
dim, norm_eps, norm_elementwise_affine
)
else:
self.attn2 = None
self.attn2_norm = None
self.norm2 = make_norm_layer(dim, norm_eps, norm_elementwise_affine)
# 3. Feed-forward
self.ff = FeedForward(
dim,
dropout=dropout,
activation_fn=activation_fn,
final_dropout=final_dropout,
inner_dim=ff_inner_dim,
bias=ff_bias,
)
# 5. Scale-shift for PixArt-Alpha.
if adaptive_norm != "none":
num_ada_params = 4 if adaptive_norm == "single_scale" else 6
self.scale_shift_table = nn.Parameter(
torch.randn(num_ada_params, dim) / dim**0.5
)
# let chunk size default to None
self._chunk_size = None
self._chunk_dim = 0
def set_use_tpu_flash_attention(self):
r"""
Function sets the flag in this object and propagates down the children. The flag will enforce the usage of TPU
attention kernel.
"""
self.use_tpu_flash_attention = True
self.attn1.set_use_tpu_flash_attention()
self.attn2.set_use_tpu_flash_attention()
def set_chunk_feed_forward(self, chunk_size: Optional[int], dim: int = 0):
# Sets chunk feed-forward
self._chunk_size = chunk_size
self._chunk_dim = dim
def forward(
self,
hidden_states: torch.FloatTensor,
freqs_cis: Optional[Tuple[torch.FloatTensor, torch.FloatTensor]] = None,
attention_mask: Optional[torch.FloatTensor] = None,
encoder_hidden_states: Optional[torch.FloatTensor] = None,
encoder_attention_mask: Optional[torch.FloatTensor] = None,
timestep: Optional[torch.LongTensor] = None,
cross_attention_kwargs: Dict[str, Any] = None,
class_labels: Optional[torch.LongTensor] = None,
skip_layer_mask: Optional[torch.Tensor] = None,
skip_layer_strategy: Optional[SkipLayerStrategy] = None,
) -> torch.FloatTensor:
if cross_attention_kwargs is not None:
if cross_attention_kwargs.get("scale", None) is not None:
logger.warning(
"Passing `scale` to `cross_attention_kwargs` is depcrecated. `scale` will be ignored."
)
# Notice that normalization is always applied before the real computation in the following blocks.
# 0. Self-Attention
batch_size = hidden_states.shape[0]
original_hidden_states = hidden_states
norm_hidden_states = self.norm1(hidden_states)
# Apply ada_norm_single
if self.adaptive_norm in ["single_scale_shift", "single_scale"]:
assert timestep.ndim == 3 # [batch, 1 or num_tokens, embedding_dim]
num_ada_params = self.scale_shift_table.shape[0]
ada_values = self.scale_shift_table[None, None] + timestep.reshape(
batch_size, timestep.shape[1], num_ada_params, -1
)
if self.adaptive_norm == "single_scale_shift":
shift_msa, scale_msa, gate_msa, shift_mlp, scale_mlp, gate_mlp = (
ada_values.unbind(dim=2)
)
norm_hidden_states = norm_hidden_states * (1 + scale_msa) + shift_msa
else:
scale_msa, gate_msa, scale_mlp, gate_mlp = ada_values.unbind(dim=2)
norm_hidden_states = norm_hidden_states * (1 + scale_msa)
elif self.adaptive_norm == "none":
scale_msa, gate_msa, scale_mlp, gate_mlp = None, None, None, None
else:
raise ValueError(f"Unknown adaptive norm type: {self.adaptive_norm}")
norm_hidden_states = norm_hidden_states.squeeze(
1
) # TODO: Check if this is needed
# 1. Prepare GLIGEN inputs
cross_attention_kwargs = (
cross_attention_kwargs.copy() if cross_attention_kwargs is not None else {}
)
attn_output = self.attn1(
norm_hidden_states,
freqs_cis=freqs_cis,
encoder_hidden_states=(
encoder_hidden_states if self.only_cross_attention else None
),
attention_mask=attention_mask,
skip_layer_mask=skip_layer_mask,
skip_layer_strategy=skip_layer_strategy,
**cross_attention_kwargs,
)
if gate_msa is not None:
attn_output = gate_msa * attn_output
hidden_states = attn_output + hidden_states
if hidden_states.ndim == 4:
hidden_states = hidden_states.squeeze(1)
# 3. Cross-Attention
if self.attn2 is not None:
if self.adaptive_norm == "none":
attn_input = self.attn2_norm(hidden_states)
else:
attn_input = hidden_states
attn_output = self.attn2(
attn_input,
freqs_cis=freqs_cis,
encoder_hidden_states=encoder_hidden_states,
attention_mask=encoder_attention_mask,
**cross_attention_kwargs,
)
hidden_states = attn_output + hidden_states
# 4. Feed-forward
norm_hidden_states = self.norm2(hidden_states)
if self.adaptive_norm == "single_scale_shift":
norm_hidden_states = norm_hidden_states * (1 + scale_mlp) + shift_mlp
elif self.adaptive_norm == "single_scale":
norm_hidden_states = norm_hidden_states * (1 + scale_mlp)
elif self.adaptive_norm == "none":
pass
else:
raise ValueError(f"Unknown adaptive norm type: {self.adaptive_norm}")
if self._chunk_size is not None:
# "feed_forward_chunk_size" can be used to save memory
ff_output = _chunked_feed_forward(
self.ff, norm_hidden_states, self._chunk_dim, self._chunk_size
)
else:
ff_output = self.ff(norm_hidden_states)
if gate_mlp is not None:
ff_output = gate_mlp * ff_output
hidden_states = ff_output + hidden_states
if hidden_states.ndim == 4:
hidden_states = hidden_states.squeeze(1)
if (
skip_layer_mask is not None
and skip_layer_strategy == SkipLayerStrategy.TransformerBlock
):
skip_layer_mask = skip_layer_mask.view(-1, 1, 1)
hidden_states = hidden_states * skip_layer_mask + original_hidden_states * (
1.0 - skip_layer_mask
)
return hidden_states
@maybe_allow_in_graph
class Attention(nn.Module):
r"""
A cross attention layer.
Parameters:
query_dim (`int`):
The number of channels in the query.
cross_attention_dim (`int`, *optional*):
The number of channels in the encoder_hidden_states. If not given, defaults to `query_dim`.
heads (`int`, *optional*, defaults to 8):
The number of heads to use for multi-head attention.
dim_head (`int`, *optional*, defaults to 64):
The number of channels in each head.
dropout (`float`, *optional*, defaults to 0.0):
The dropout probability to use.
bias (`bool`, *optional*, defaults to False):
Set to `True` for the query, key, and value linear layers to contain a bias parameter.
upcast_attention (`bool`, *optional*, defaults to False):
Set to `True` to upcast the attention computation to `float32`.
upcast_softmax (`bool`, *optional*, defaults to False):
Set to `True` to upcast the softmax computation to `float32`.
cross_attention_norm (`str`, *optional*, defaults to `None`):
The type of normalization to use for the cross attention. Can be `None`, `layer_norm`, or `group_norm`.
cross_attention_norm_num_groups (`int`, *optional*, defaults to 32):
The number of groups to use for the group norm in the cross attention.
added_kv_proj_dim (`int`, *optional*, defaults to `None`):
The number of channels to use for the added key and value projections. If `None`, no projection is used.
norm_num_groups (`int`, *optional*, defaults to `None`):
The number of groups to use for the group norm in the attention.
spatial_norm_dim (`int`, *optional*, defaults to `None`):
The number of channels to use for the spatial normalization.
out_bias (`bool`, *optional*, defaults to `True`):
Set to `True` to use a bias in the output linear layer.
scale_qk (`bool`, *optional*, defaults to `True`):
Set to `True` to scale the query and key by `1 / sqrt(dim_head)`.
qk_norm (`str`, *optional*, defaults to None):
Set to 'layer_norm' or `rms_norm` to perform query and key normalization.
only_cross_attention (`bool`, *optional*, defaults to `False`):
Set to `True` to only use cross attention and not added_kv_proj_dim. Can only be set to `True` if
`added_kv_proj_dim` is not `None`.
eps (`float`, *optional*, defaults to 1e-5):
An additional value added to the denominator in group normalization that is used for numerical stability.
rescale_output_factor (`float`, *optional*, defaults to 1.0):
A factor to rescale the output by dividing it with this value.
residual_connection (`bool`, *optional*, defaults to `False`):
Set to `True` to add the residual connection to the output.
_from_deprecated_attn_block (`bool`, *optional*, defaults to `False`):
Set to `True` if the attention block is loaded from a deprecated state dict.
processor (`AttnProcessor`, *optional*, defaults to `None`):
The attention processor to use. If `None`, defaults to `AttnProcessor2_0` if `torch 2.x` is used and
`AttnProcessor` otherwise.
"""
def __init__(
self,
query_dim: int,
cross_attention_dim: Optional[int] = None,
heads: int = 8,
dim_head: int = 64,
dropout: float = 0.0,
bias: bool = False,
upcast_attention: bool = False,
upcast_softmax: bool = False,
cross_attention_norm: Optional[str] = None,
cross_attention_norm_num_groups: int = 32,
added_kv_proj_dim: Optional[int] = None,
norm_num_groups: Optional[int] = None,
spatial_norm_dim: Optional[int] = None,
out_bias: bool = True,
scale_qk: bool = True,
qk_norm: Optional[str] = None,
only_cross_attention: bool = False,
eps: float = 1e-5,
rescale_output_factor: float = 1.0,
residual_connection: bool = False,
_from_deprecated_attn_block: bool = False,
processor: Optional["AttnProcessor"] = None,
out_dim: int = None,
use_tpu_flash_attention: bool = False,
use_rope: bool = False,
):
super().__init__()
self.inner_dim = out_dim if out_dim is not None else dim_head * heads
self.query_dim = query_dim
self.use_bias = bias
self.is_cross_attention = cross_attention_dim is not None
self.cross_attention_dim = (
cross_attention_dim if cross_attention_dim is not None else query_dim
)
self.upcast_attention = upcast_attention
self.upcast_softmax = upcast_softmax
self.rescale_output_factor = rescale_output_factor
self.residual_connection = residual_connection
self.dropout = dropout
self.fused_projections = False
self.out_dim = out_dim if out_dim is not None else query_dim
self.use_tpu_flash_attention = use_tpu_flash_attention
self.use_rope = use_rope
# we make use of this private variable to know whether this class is loaded
# with an deprecated state dict so that we can convert it on the fly
self._from_deprecated_attn_block = _from_deprecated_attn_block
self.scale_qk = scale_qk
self.scale = dim_head**-0.5 if self.scale_qk else 1.0
if qk_norm is None:
self.q_norm = nn.Identity()
self.k_norm = nn.Identity()
elif qk_norm == "rms_norm":
self.q_norm = RMSNorm(dim_head * heads, eps=1e-5)
self.k_norm = RMSNorm(dim_head * heads, eps=1e-5)
elif qk_norm == "layer_norm":
self.q_norm = nn.LayerNorm(dim_head * heads, eps=1e-5)
self.k_norm = nn.LayerNorm(dim_head * heads, eps=1e-5)
else:
raise ValueError(f"Unsupported qk_norm method: {qk_norm}")
self.heads = out_dim // dim_head if out_dim is not None else heads
# for slice_size > 0 the attention score computation
# is split across the batch axis to save memory
# You can set slice_size with `set_attention_slice`
self.sliceable_head_dim = heads
self.added_kv_proj_dim = added_kv_proj_dim
self.only_cross_attention = only_cross_attention
if self.added_kv_proj_dim is None and self.only_cross_attention:
raise ValueError(
"`only_cross_attention` can only be set to True if `added_kv_proj_dim` is not None. Make sure to set either `only_cross_attention=False` or define `added_kv_proj_dim`."
)
if norm_num_groups is not None:
self.group_norm = nn.GroupNorm(
num_channels=query_dim, num_groups=norm_num_groups, eps=eps, affine=True
)
else:
self.group_norm = None
if spatial_norm_dim is not None:
self.spatial_norm = SpatialNorm(
f_channels=query_dim, zq_channels=spatial_norm_dim
)
else:
self.spatial_norm = None
if cross_attention_norm is None:
self.norm_cross = None
elif cross_attention_norm == "layer_norm":
self.norm_cross = nn.LayerNorm(self.cross_attention_dim)
elif cross_attention_norm == "group_norm":
if self.added_kv_proj_dim is not None:
# The given `encoder_hidden_states` are initially of shape
# (batch_size, seq_len, added_kv_proj_dim) before being projected
# to (batch_size, seq_len, cross_attention_dim). The norm is applied
# before the projection, so we need to use `added_kv_proj_dim` as
# the number of channels for the group norm.
norm_cross_num_channels = added_kv_proj_dim
else:
norm_cross_num_channels = self.cross_attention_dim
self.norm_cross = nn.GroupNorm(
num_channels=norm_cross_num_channels,
num_groups=cross_attention_norm_num_groups,
eps=1e-5,
affine=True,
)
else:
raise ValueError(
f"unknown cross_attention_norm: {cross_attention_norm}. Should be None, 'layer_norm' or 'group_norm'"
)
linear_cls = nn.Linear
self.linear_cls = linear_cls
self.to_q = linear_cls(query_dim, self.inner_dim, bias=bias)
if not self.only_cross_attention:
# only relevant for the `AddedKVProcessor` classes
self.to_k = linear_cls(self.cross_attention_dim, self.inner_dim, bias=bias)
self.to_v = linear_cls(self.cross_attention_dim, self.inner_dim, bias=bias)
else:
self.to_k = None
self.to_v = None
if self.added_kv_proj_dim is not None:
self.add_k_proj = linear_cls(added_kv_proj_dim, self.inner_dim)
self.add_v_proj = linear_cls(added_kv_proj_dim, self.inner_dim)
self.to_out = nn.ModuleList([])
self.to_out.append(linear_cls(self.inner_dim, self.out_dim, bias=out_bias))
self.to_out.append(nn.Dropout(dropout))
# set attention processor
# We use the AttnProcessor2_0 by default when torch 2.x is used which uses
# torch.nn.functional.scaled_dot_product_attention for native Flash/memory_efficient_attention
# but only if it has the default `scale` argument. TODO remove scale_qk check when we move to torch 2.1
if processor is None:
processor = AttnProcessor2_0()
self.set_processor(processor)
def set_use_tpu_flash_attention(self):
r"""
Function sets the flag in this object. The flag will enforce the usage of TPU attention kernel.
"""
self.use_tpu_flash_attention = True
def set_processor(self, processor: "AttnProcessor") -> None:
r"""
Set the attention processor to use.
Args:
processor (`AttnProcessor`):
The attention processor to use.
"""
# if current processor is in `self._modules` and if passed `processor` is not, we need to
# pop `processor` from `self._modules`
if (
hasattr(self, "processor")
and isinstance(self.processor, torch.nn.Module)
and not isinstance(processor, torch.nn.Module)
):
logger.info(
f"You are removing possibly trained weights of {self.processor} with {processor}"
)
self._modules.pop("processor")
self.processor = processor
def get_processor(
self, return_deprecated_lora: bool = False
) -> "AttentionProcessor": # noqa: F821
r"""
Get the attention processor in use.
Args:
return_deprecated_lora (`bool`, *optional*, defaults to `False`):
Set to `True` to return the deprecated LoRA attention processor.
Returns:
"AttentionProcessor": The attention processor in use.
"""
if not return_deprecated_lora:
return self.processor
# TODO(Sayak, Patrick). The rest of the function is needed to ensure backwards compatible
# serialization format for LoRA Attention Processors. It should be deleted once the integration
# with PEFT is completed.
is_lora_activated = {
name: module.lora_layer is not None
for name, module in self.named_modules()
if hasattr(module, "lora_layer")
}
# 1. if no layer has a LoRA activated we can return the processor as usual
if not any(is_lora_activated.values()):
return self.processor
# If doesn't apply LoRA do `add_k_proj` or `add_v_proj`
is_lora_activated.pop("add_k_proj", None)
is_lora_activated.pop("add_v_proj", None)
# 2. else it is not posssible that only some layers have LoRA activated
if not all(is_lora_activated.values()):
raise ValueError(
f"Make sure that either all layers or no layers have LoRA activated, but have {is_lora_activated}"
)
# 3. And we need to merge the current LoRA layers into the corresponding LoRA attention processor
non_lora_processor_cls_name = self.processor.__class__.__name__
lora_processor_cls = getattr(
import_module(__name__), "LoRA" + non_lora_processor_cls_name
)
hidden_size = self.inner_dim
# now create a LoRA attention processor from the LoRA layers
if lora_processor_cls in [
LoRAAttnProcessor,
LoRAAttnProcessor2_0,
LoRAXFormersAttnProcessor,
]:
kwargs = {
"cross_attention_dim": self.cross_attention_dim,
"rank": self.to_q.lora_layer.rank,
"network_alpha": self.to_q.lora_layer.network_alpha,
"q_rank": self.to_q.lora_layer.rank,
"q_hidden_size": self.to_q.lora_layer.out_features,
"k_rank": self.to_k.lora_layer.rank,
"k_hidden_size": self.to_k.lora_layer.out_features,
"v_rank": self.to_v.lora_layer.rank,
"v_hidden_size": self.to_v.lora_layer.out_features,
"out_rank": self.to_out[0].lora_layer.rank,
"out_hidden_size": self.to_out[0].lora_layer.out_features,
}
if hasattr(self.processor, "attention_op"):
kwargs["attention_op"] = self.processor.attention_op
lora_processor = lora_processor_cls(hidden_size, **kwargs)
lora_processor.to_q_lora.load_state_dict(self.to_q.lora_layer.state_dict())
lora_processor.to_k_lora.load_state_dict(self.to_k.lora_layer.state_dict())
lora_processor.to_v_lora.load_state_dict(self.to_v.lora_layer.state_dict())
lora_processor.to_out_lora.load_state_dict(
self.to_out[0].lora_layer.state_dict()
)
elif lora_processor_cls == LoRAAttnAddedKVProcessor:
lora_processor = lora_processor_cls(
hidden_size,
cross_attention_dim=self.add_k_proj.weight.shape[0],
rank=self.to_q.lora_layer.rank,
network_alpha=self.to_q.lora_layer.network_alpha,
)
lora_processor.to_q_lora.load_state_dict(self.to_q.lora_layer.state_dict())
lora_processor.to_k_lora.load_state_dict(self.to_k.lora_layer.state_dict())
lora_processor.to_v_lora.load_state_dict(self.to_v.lora_layer.state_dict())
lora_processor.to_out_lora.load_state_dict(
self.to_out[0].lora_layer.state_dict()
)
# only save if used
if self.add_k_proj.lora_layer is not None:
lora_processor.add_k_proj_lora.load_state_dict(
self.add_k_proj.lora_layer.state_dict()
)
lora_processor.add_v_proj_lora.load_state_dict(
self.add_v_proj.lora_layer.state_dict()
)
else:
lora_processor.add_k_proj_lora = None
lora_processor.add_v_proj_lora = None
else:
raise ValueError(f"{lora_processor_cls} does not exist.")
return lora_processor
def forward(
self,
hidden_states: torch.FloatTensor,
freqs_cis: Optional[Tuple[torch.FloatTensor, torch.FloatTensor]] = None,
encoder_hidden_states: Optional[torch.FloatTensor] = None,
attention_mask: Optional[torch.FloatTensor] = None,
skip_layer_mask: Optional[torch.Tensor] = None,
skip_layer_strategy: Optional[SkipLayerStrategy] = None,
**cross_attention_kwargs,
) -> torch.Tensor:
r"""
The forward method of the `Attention` class.
Args:
hidden_states (`torch.Tensor`):
The hidden states of the query.
encoder_hidden_states (`torch.Tensor`, *optional*):
The hidden states of the encoder.
attention_mask (`torch.Tensor`, *optional*):
The attention mask to use. If `None`, no mask is applied.
skip_layer_mask (`torch.Tensor`, *optional*):
The skip layer mask to use. If `None`, no mask is applied.
skip_layer_strategy (`SkipLayerStrategy`, *optional*, defaults to `None`):
Controls which layers to skip for spatiotemporal guidance.
**cross_attention_kwargs:
Additional keyword arguments to pass along to the cross attention.
Returns:
`torch.Tensor`: The output of the attention layer.
"""
# The `Attention` class can call different attention processors / attention functions
# here we simply pass along all tensors to the selected processor class
# For standard processors that are defined here, `**cross_attention_kwargs` is empty
attn_parameters = set(
inspect.signature(self.processor.__call__).parameters.keys()
)
unused_kwargs = [
k for k, _ in cross_attention_kwargs.items() if k not in attn_parameters
]
if len(unused_kwargs) > 0:
logger.warning(
f"cross_attention_kwargs {unused_kwargs} are not expected by"
f" {self.processor.__class__.__name__} and will be ignored."
)
cross_attention_kwargs = {
k: w for k, w in cross_attention_kwargs.items() if k in attn_parameters
}
return self.processor(
self,
hidden_states,
freqs_cis=freqs_cis,
encoder_hidden_states=encoder_hidden_states,
attention_mask=attention_mask,
skip_layer_mask=skip_layer_mask,
skip_layer_strategy=skip_layer_strategy,
**cross_attention_kwargs,
)
def batch_to_head_dim(self, tensor: torch.Tensor) -> torch.Tensor:
r"""
Reshape the tensor from `[batch_size, seq_len, dim]` to `[batch_size // heads, seq_len, dim * heads]`. `heads`
is the number of heads initialized while constructing the `Attention` class.
Args:
tensor (`torch.Tensor`): The tensor to reshape.
Returns:
`torch.Tensor`: The reshaped tensor.
"""
head_size = self.heads
batch_size, seq_len, dim = tensor.shape
tensor = tensor.reshape(batch_size // head_size, head_size, seq_len, dim)
tensor = tensor.permute(0, 2, 1, 3).reshape(
batch_size // head_size, seq_len, dim * head_size
)
return tensor
def head_to_batch_dim(self, tensor: torch.Tensor, out_dim: int = 3) -> torch.Tensor:
r"""
Reshape the tensor from `[batch_size, seq_len, dim]` to `[batch_size, seq_len, heads, dim // heads]` `heads` is
the number of heads initialized while constructing the `Attention` class.
Args:
tensor (`torch.Tensor`): The tensor to reshape.
out_dim (`int`, *optional*, defaults to `3`): The output dimension of the tensor. If `3`, the tensor is
reshaped to `[batch_size * heads, seq_len, dim // heads]`.
Returns:
`torch.Tensor`: The reshaped tensor.
"""
head_size = self.heads
if tensor.ndim == 3:
batch_size, seq_len, dim = tensor.shape
extra_dim = 1
else:
batch_size, extra_dim, seq_len, dim = tensor.shape
tensor = tensor.reshape(
batch_size, seq_len * extra_dim, head_size, dim // head_size
)
tensor = tensor.permute(0, 2, 1, 3)
if out_dim == 3:
tensor = tensor.reshape(
batch_size * head_size, seq_len * extra_dim, dim // head_size
)
return tensor
def get_attention_scores(
self,
query: torch.Tensor,
key: torch.Tensor,
attention_mask: torch.Tensor = None,
) -> torch.Tensor:
r"""
Compute the attention scores.
Args:
query (`torch.Tensor`): The query tensor.
key (`torch.Tensor`): The key tensor.
attention_mask (`torch.Tensor`, *optional*): The attention mask to use. If `None`, no mask is applied.
Returns:
`torch.Tensor`: The attention probabilities/scores.
"""
dtype = query.dtype
if self.upcast_attention:
query = query.float()
key = key.float()
if attention_mask is None:
baddbmm_input = torch.empty(
query.shape[0],
query.shape[1],
key.shape[1],
dtype=query.dtype,
device=query.device,
)
beta = 0
else:
baddbmm_input = attention_mask
beta = 1
attention_scores = torch.baddbmm(
baddbmm_input,
query,
key.transpose(-1, -2),
beta=beta,
alpha=self.scale,
)
del baddbmm_input
if self.upcast_softmax:
attention_scores = attention_scores.float()
attention_probs = attention_scores.softmax(dim=-1)
del attention_scores
attention_probs = attention_probs.to(dtype)
return attention_probs
def prepare_attention_mask(
self,
attention_mask: torch.Tensor,
target_length: int,
batch_size: int,
out_dim: int = 3,
) -> torch.Tensor:
r"""
Prepare the attention mask for the attention computation.
Args:
attention_mask (`torch.Tensor`):
The attention mask to prepare.
target_length (`int`):
The target length of the attention mask. This is the length of the attention mask after padding.
batch_size (`int`):
The batch size, which is used to repeat the attention mask.
out_dim (`int`, *optional*, defaults to `3`):
The output dimension of the attention mask. Can be either `3` or `4`.
Returns:
`torch.Tensor`: The prepared attention mask.
"""
head_size = self.heads
if attention_mask is None:
return attention_mask
current_length: int = attention_mask.shape[-1]
if current_length != target_length:
if attention_mask.device.type == "mps":
# HACK: MPS: Does not support padding by greater than dimension of input tensor.
# Instead, we can manually construct the padding tensor.
padding_shape = (
attention_mask.shape[0],
attention_mask.shape[1],
target_length,
)
padding = torch.zeros(
padding_shape,
dtype=attention_mask.dtype,
device=attention_mask.device,
)
attention_mask = torch.cat([attention_mask, padding], dim=2)
else:
# TODO: for pipelines such as stable-diffusion, padding cross-attn mask:
# we want to instead pad by (0, remaining_length), where remaining_length is:
# remaining_length: int = target_length - current_length
# TODO: re-enable tests/models/test_models_unet_2d_condition.py#test_model_xattn_padding
attention_mask = F.pad(attention_mask, (0, target_length), value=0.0)
if out_dim == 3:
if attention_mask.shape[0] < batch_size * head_size:
attention_mask = attention_mask.repeat_interleave(head_size, dim=0)
elif out_dim == 4:
attention_mask = attention_mask.unsqueeze(1)
attention_mask = attention_mask.repeat_interleave(head_size, dim=1)
return attention_mask
def norm_encoder_hidden_states(
self, encoder_hidden_states: torch.Tensor
) -> torch.Tensor:
r"""
Normalize the encoder hidden states. Requires `self.norm_cross` to be specified when constructing the
`Attention` class.
Args:
encoder_hidden_states (`torch.Tensor`): Hidden states of the encoder.
Returns:
`torch.Tensor`: The normalized encoder hidden states.
"""
assert (
self.norm_cross is not None
), "self.norm_cross must be defined to call self.norm_encoder_hidden_states"
if isinstance(self.norm_cross, nn.LayerNorm):
encoder_hidden_states = self.norm_cross(encoder_hidden_states)
elif isinstance(self.norm_cross, nn.GroupNorm):
# Group norm norms along the channels dimension and expects
# input to be in the shape of (N, C, *). In this case, we want
# to norm along the hidden dimension, so we need to move
# (batch_size, sequence_length, hidden_size) ->
# (batch_size, hidden_size, sequence_length)
encoder_hidden_states = encoder_hidden_states.transpose(1, 2)
encoder_hidden_states = self.norm_cross(encoder_hidden_states)
encoder_hidden_states = encoder_hidden_states.transpose(1, 2)
else:
assert False
return encoder_hidden_states
@staticmethod
def apply_rotary_emb(
input_tensor: torch.Tensor,
freqs_cis: Tuple[torch.FloatTensor, torch.FloatTensor],
) -> Tuple[torch.Tensor, torch.Tensor]:
cos_freqs = freqs_cis[0]
sin_freqs = freqs_cis[1]
t_dup = rearrange(input_tensor, "... (d r) -> ... d r", r=2)
t1, t2 = t_dup.unbind(dim=-1)
t_dup = torch.stack((-t2, t1), dim=-1)
input_tensor_rot = rearrange(t_dup, "... d r -> ... (d r)")
out = input_tensor * cos_freqs + input_tensor_rot * sin_freqs
return out
class AttnProcessor2_0:
r"""
Processor for implementing scaled dot-product attention (enabled by default if you're using PyTorch 2.0).
"""
def __init__(self):
pass
def __call__(
self,
attn: Attention,
hidden_states: torch.FloatTensor,
freqs_cis: Tuple[torch.FloatTensor, torch.FloatTensor],
encoder_hidden_states: Optional[torch.FloatTensor] = None,
attention_mask: Optional[torch.FloatTensor] = None,
temb: Optional[torch.FloatTensor] = None,
skip_layer_mask: Optional[torch.FloatTensor] = None,
skip_layer_strategy: Optional[SkipLayerStrategy] = None,
*args,
**kwargs,
) -> torch.FloatTensor:
if len(args) > 0 or kwargs.get("scale", None) is not None:
deprecation_message = "The `scale` argument is deprecated and will be ignored. Please remove it, as passing it will raise an error in the future. `scale` should directly be passed while calling the underlying pipeline component i.e., via `cross_attention_kwargs`."
deprecate("scale", "1.0.0", deprecation_message)
residual = hidden_states
if attn.spatial_norm is not None:
hidden_states = attn.spatial_norm(hidden_states, temb)
input_ndim = hidden_states.ndim
if input_ndim == 4:
batch_size, channel, height, width = hidden_states.shape
hidden_states = hidden_states.view(
batch_size, channel, height * width
).transpose(1, 2)
batch_size, sequence_length, _ = (
hidden_states.shape
if encoder_hidden_states is None
else encoder_hidden_states.shape
)
if skip_layer_mask is not None:
skip_layer_mask = skip_layer_mask.reshape(batch_size, 1, 1)
if (attention_mask is not None) and (not attn.use_tpu_flash_attention):
attention_mask = attn.prepare_attention_mask(
attention_mask, sequence_length, batch_size
)
# scaled_dot_product_attention expects attention_mask shape to be
# (batch, heads, source_length, target_length)
attention_mask = attention_mask.view(
batch_size, attn.heads, -1, attention_mask.shape[-1]
)
if attn.group_norm is not None:
hidden_states = attn.group_norm(hidden_states.transpose(1, 2)).transpose(
1, 2
)
query = attn.to_q(hidden_states)
query = attn.q_norm(query)
if encoder_hidden_states is not None:
if attn.norm_cross:
encoder_hidden_states = attn.norm_encoder_hidden_states(
encoder_hidden_states
)
key = attn.to_k(encoder_hidden_states)
key = attn.k_norm(key)
else: # if no context provided do self-attention
encoder_hidden_states = hidden_states
key = attn.to_k(hidden_states)
key = attn.k_norm(key)
if attn.use_rope:
key = attn.apply_rotary_emb(key, freqs_cis)
query = attn.apply_rotary_emb(query, freqs_cis)
value = attn.to_v(encoder_hidden_states)
value_for_stg = value
inner_dim = key.shape[-1]
head_dim = inner_dim // attn.heads
query = query.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
key = key.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
value = value.view(batch_size, -1, attn.heads, head_dim).transpose(1, 2)
# the output of sdp = (batch, num_heads, seq_len, head_dim)
if attn.use_tpu_flash_attention: # use tpu attention offload 'flash attention'
q_segment_indexes = None
if (
attention_mask is not None
): # if mask is required need to tune both segmenIds fields
# attention_mask = torch.squeeze(attention_mask).to(torch.float32)
attention_mask = attention_mask.to(torch.float32)
q_segment_indexes = torch.ones(
batch_size, query.shape[2], device=query.device, dtype=torch.float32
)
assert (
attention_mask.shape[1] == key.shape[2]
), f"ERROR: KEY SHAPE must be same as attention mask [{key.shape[2]}, {attention_mask.shape[1]}]"
assert (
query.shape[2] % 128 == 0
), f"ERROR: QUERY SHAPE must be divisible by 128 (TPU limitation) [{query.shape[2]}]"
assert (
key.shape[2] % 128 == 0
), f"ERROR: KEY SHAPE must be divisible by 128 (TPU limitation) [{key.shape[2]}]"
# run the TPU kernel implemented in jax with pallas
hidden_states_a = flash_attention(
q=query,
k=key,
v=value,
q_segment_ids=q_segment_indexes,
kv_segment_ids=attention_mask,
sm_scale=attn.scale,
)
else:
hidden_states_a = F.scaled_dot_product_attention(
query,
key,
value,
attn_mask=attention_mask,
dropout_p=0.0,
is_causal=False,
)
hidden_states_a = hidden_states_a.transpose(1, 2).reshape(
batch_size, -1, attn.heads * head_dim
)
hidden_states_a = hidden_states_a.to(query.dtype)
if (
skip_layer_mask is not None
and skip_layer_strategy == SkipLayerStrategy.AttentionSkip
):
hidden_states = hidden_states_a * skip_layer_mask + hidden_states * (
1.0 - skip_layer_mask
)
elif (
skip_layer_mask is not None
and skip_layer_strategy == SkipLayerStrategy.AttentionValues
):
hidden_states = hidden_states_a * skip_layer_mask + value_for_stg * (
1.0 - skip_layer_mask
)
else:
hidden_states = hidden_states_a
# linear proj
hidden_states = attn.to_out[0](hidden_states)
# dropout
hidden_states = attn.to_out[1](hidden_states)
if input_ndim == 4:
hidden_states = hidden_states.transpose(-1, -2).reshape(
batch_size, channel, height, width
)
if (
skip_layer_mask is not None
and skip_layer_strategy == SkipLayerStrategy.Residual
):
skip_layer_mask = skip_layer_mask.reshape(batch_size, 1, 1, 1)
if attn.residual_connection:
if (
skip_layer_mask is not None
and skip_layer_strategy == SkipLayerStrategy.Residual
):
hidden_states = hidden_states + residual * skip_layer_mask
else:
hidden_states = hidden_states + residual
hidden_states = hidden_states / attn.rescale_output_factor
return hidden_states
class AttnProcessor:
r"""
Default processor for performing attention-related computations.
"""
def __call__(
self,
attn: Attention,
hidden_states: torch.FloatTensor,
encoder_hidden_states: Optional[torch.FloatTensor] = None,
attention_mask: Optional[torch.FloatTensor] = None,
temb: Optional[torch.FloatTensor] = None,
*args,
**kwargs,
) -> torch.Tensor:
if len(args) > 0 or kwargs.get("scale", None) is not None:
deprecation_message = "The `scale` argument is deprecated and will be ignored. Please remove it, as passing it will raise an error in the future. `scale` should directly be passed while calling the underlying pipeline component i.e., via `cross_attention_kwargs`."
deprecate("scale", "1.0.0", deprecation_message)
residual = hidden_states
if attn.spatial_norm is not None:
hidden_states = attn.spatial_norm(hidden_states, temb)
input_ndim = hidden_states.ndim
if input_ndim == 4:
batch_size, channel, height, width = hidden_states.shape
hidden_states = hidden_states.view(
batch_size, channel, height * width
).transpose(1, 2)
batch_size, sequence_length, _ = (
hidden_states.shape
if encoder_hidden_states is None
else encoder_hidden_states.shape
)
attention_mask = attn.prepare_attention_mask(
attention_mask, sequence_length, batch_size
)
if attn.group_norm is not None:
hidden_states = attn.group_norm(hidden_states.transpose(1, 2)).transpose(
1, 2
)
query = attn.to_q(hidden_states)
if encoder_hidden_states is None:
encoder_hidden_states = hidden_states
elif attn.norm_cross:
encoder_hidden_states = attn.norm_encoder_hidden_states(
encoder_hidden_states
)
key = attn.to_k(encoder_hidden_states)
value = attn.to_v(encoder_hidden_states)
query = attn.head_to_batch_dim(query)
key = attn.head_to_batch_dim(key)
value = attn.head_to_batch_dim(value)
query = attn.q_norm(query)
key = attn.k_norm(key)
attention_probs = attn.get_attention_scores(query, key, attention_mask)
hidden_states = torch.bmm(attention_probs, value)
hidden_states = attn.batch_to_head_dim(hidden_states)
# linear proj
hidden_states = attn.to_out[0](hidden_states)
# dropout
hidden_states = attn.to_out[1](hidden_states)
if input_ndim == 4:
hidden_states = hidden_states.transpose(-1, -2).reshape(
batch_size, channel, height, width
)
if attn.residual_connection:
hidden_states = hidden_states + residual
hidden_states = hidden_states / attn.rescale_output_factor
return hidden_states
class FeedForward(nn.Module):
r"""
A feed-forward layer.
Parameters:
dim (`int`): The number of channels in the input.
dim_out (`int`, *optional*): The number of channels in the output. If not given, defaults to `dim`.
mult (`int`, *optional*, defaults to 4): The multiplier to use for the hidden dimension.
dropout (`float`, *optional*, defaults to 0.0): The dropout probability to use.
activation_fn (`str`, *optional*, defaults to `"geglu"`): Activation function to be used in feed-forward.
final_dropout (`bool` *optional*, defaults to False): Apply a final dropout.
bias (`bool`, defaults to True): Whether to use a bias in the linear layer.
"""
def __init__(
self,
dim: int,
dim_out: Optional[int] = None,
mult: int = 4,
dropout: float = 0.0,
activation_fn: str = "geglu",
final_dropout: bool = False,
inner_dim=None,
bias: bool = True,
):
super().__init__()
if inner_dim is None:
inner_dim = int(dim * mult)
dim_out = dim_out if dim_out is not None else dim
linear_cls = nn.Linear
if activation_fn == "gelu":
act_fn = GELU(dim, inner_dim, bias=bias)
elif activation_fn == "gelu-approximate":
act_fn = GELU(dim, inner_dim, approximate="tanh", bias=bias)
elif activation_fn == "geglu":
act_fn = GEGLU(dim, inner_dim, bias=bias)
elif activation_fn == "geglu-approximate":
act_fn = ApproximateGELU(dim, inner_dim, bias=bias)
else:
raise ValueError(f"Unsupported activation function: {activation_fn}")
self.net = nn.ModuleList([])
# project in
self.net.append(act_fn)
# project dropout
self.net.append(nn.Dropout(dropout))
# project out
self.net.append(linear_cls(inner_dim, dim_out, bias=bias))
# FF as used in Vision Transformer, MLP-Mixer, etc. have a final dropout
if final_dropout:
self.net.append(nn.Dropout(dropout))
def forward(self, hidden_states: torch.Tensor, scale: float = 1.0) -> torch.Tensor:
compatible_cls = (GEGLU, LoRACompatibleLinear)
for module in self.net:
if isinstance(module, compatible_cls):
hidden_states = module(hidden_states, scale)
else:
hidden_states = module(hidden_states)
return hidden_states
================================================
FILE: ltx_video/models/transformers/embeddings.py
================================================
# Adapted from: https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/embeddings.py
import math
import numpy as np
import torch
from einops import rearrange
from torch import nn
def get_timestep_embedding(
timesteps: torch.Tensor,
embedding_dim: int,
flip_sin_to_cos: bool = False,
downscale_freq_shift: float = 1,
scale: float = 1,
max_period: int = 10000,
):
"""
This matches the implementation in Denoising Diffusion Probabilistic Models: Create sinusoidal timestep embeddings.
:param timesteps: a 1-D Tensor of N indices, one per batch element.
These may be fractional.
:param embedding_dim: the dimension of the output. :param max_period: controls the minimum frequency of the
embeddings. :return: an [N x dim] Tensor of positional embeddings.
"""
assert len(timesteps.shape) == 1, "Timesteps should be a 1d-array"
half_dim = embedding_dim // 2
exponent = -math.log(max_period) * torch.arange(
start=0, end=half_dim, dtype=torch.float32, device=timesteps.device
)
exponent = exponent / (half_dim - downscale_freq_shift)
emb = torch.exp(exponent)
emb = timesteps[:, None].float() * emb[None, :]
# scale embeddings
emb = scale * emb
# concat sine and cosine embeddings
emb = torch.cat([torch.sin(emb), torch.cos(emb)], dim=-1)
# flip sine and cosine embeddings
if flip_sin_to_cos:
emb = torch.cat([emb[:, half_dim:], emb[:, :half_dim]], dim=-1)
# zero pad
if embedding_dim % 2 == 1:
emb = torch.nn.functional.pad(emb, (0, 1, 0, 0))
return emb
def get_3d_sincos_pos_embed(embed_dim, grid, w, h, f):
"""
grid_size: int of the grid height and width return: pos_embed: [grid_size*grid_size, embed_dim] or
[1+grid_size*grid_size, embed_dim] (w/ or w/o cls_token)
"""
grid = rearrange(grid, "c (f h w) -> c f h w", h=h, w=w)
grid = rearrange(grid, "c f h w -> c h w f", h=h, w=w)
grid = grid.reshape([3, 1, w, h, f])
pos_embed = get_3d_sincos_pos_embed_from_grid(embed_dim, grid)
pos_embed = pos_embed.transpose(1, 0, 2, 3)
return rearrange(pos_embed, "h w f c -> (f h w) c")
def get_3d_sincos_pos_embed_from_grid(embed_dim, grid):
if embed_dim % 3 != 0:
raise ValueError("embed_dim must be divisible by 3")
# use half of dimensions to encode grid_h
emb_f = get_1d_sincos_pos_embed_from_grid(embed_dim // 3, grid[0]) # (H*W*T, D/3)
emb_h = get_1d_sincos_pos_embed_from_grid(embed_dim // 3, grid[1]) # (H*W*T, D/3)
emb_w = get_1d_sincos_pos_embed_from_grid(embed_dim // 3, grid[2]) # (H*W*T, D/3)
emb = np.concatenate([emb_h, emb_w, emb_f], axis=-1) # (H*W*T, D)
return emb
def get_1d_sincos_pos_embed_from_grid(embed_dim, pos):
"""
embed_dim: output dimension for each position pos: a list of positions to be encoded: size (M,) out: (M, D)
"""
if embed_dim % 2 != 0:
raise ValueError("embed_dim must be divisible by 2")
omega = np.arange(embed_dim // 2, dtype=np.float64)
omega /= embed_dim / 2.0
omega = 1.0 / 10000**omega # (D/2,)
pos_shape = pos.shape
pos = pos.reshape(-1)
out = np.einsum("m,d->md", pos, omega) # (M, D/2), outer product
out = out.reshape([*pos_shape, -1])[0]
emb_sin = np.sin(out) # (M, D/2)
emb_cos = np.cos(out) # (M, D/2)
emb = np.concatenate([emb_sin, emb_cos], axis=-1) # (M, D)
return emb
class SinusoidalPositionalEmbedding(nn.Module):
"""Apply positional information to a sequence of embeddings.
Takes in a sequence of embeddings with shape (batch_size, seq_length, embed_dim) and adds positional embeddings to
them
Args:
embed_dim: (int): Dimension of the positional embedding.
max_seq_length: Maximum sequence length to apply positional embeddings
"""
def __init__(self, embed_dim: int, max_seq_length: int = 32):
super().__init__()
position = torch.arange(max_seq_length).unsqueeze(1)
div_term = torch.exp(
torch.arange(0, embed_dim, 2) * (-math.log(10000.0) / embed_dim)
)
pe = torch.zeros(1, max_seq_length, embed_dim)
pe[0, :, 0::2] = torch.sin(position * div_term)
pe[0, :, 1::2] = torch.cos(position * div_term)
self.register_buffer("pe", pe)
def forward(self, x):
_, seq_length, _ = x.shape
x = x + self.pe[:, :seq_length]
return x
================================================
FILE: ltx_video/models/transformers/symmetric_patchifier.py
================================================
from abc import ABC, abstractmethod
from typing import Tuple
import torch
from diffusers.configuration_utils import ConfigMixin
from einops import rearrange
from torch import Tensor
class Patchifier(ConfigMixin, ABC):
def __init__(self, patch_size: int):
super().__init__()
self._patch_size = (1, patch_size, patch_size)
@abstractmethod
def patchify(self, latents: Tensor) -> Tuple[Tensor, Tensor]:
raise NotImplementedError("Patchify method not implemented")
@abstractmethod
def unpatchify(
self,
latents: Tensor,
output_height: int,
output_width: int,
out_channels: int,
) -> Tuple[Tensor, Tensor]:
pass
@property
def patch_size(self):
return self._patch_size
def get_latent_coords(
self, latent_num_frames, latent_height, latent_width, batch_size, device
):
"""
Return a tensor of shape [batch_size, 3, num_patches] containing the
top-left corner latent coordinates of each latent patch.
The tensor is repeated for each batch element.
"""
latent_sample_coords = torch.meshgrid(
torch.arange(0, latent_num_frames, self._patch_size[0], device=device),
torch.arange(0, latent_height, self._patch_size[1], device=device),
torch.arange(0, latent_width, self._patch_size[2], device=device),
)
latent_sample_coords = torch.stack(latent_sample_coords, dim=0)
latent_coords = latent_sample_coords.unsqueeze(0).repeat(batch_size, 1, 1, 1, 1)
latent_coords = rearrange(
latent_coords, "b c f h w -> b c (f h w)", b=batch_size
)
return latent_coords
class SymmetricPatchifier(Patchifier):
def patchify(self, latents: Tensor) -> Tuple[Tensor, Tensor]:
b, _, f, h, w = latents.shape
latent_coords = self.get_latent_coords(f, h, w, b, latents.device)
latents = rearrange(
latents,
"b c (f p1) (h p2) (w p3) -> b (f h w) (c p1 p2 p3)",
p1=self._patch_size[0],
p2=self._patch_size[1],
p3=self._patch_size[2],
)
return latents, latent_coords
def unpatchify(
self,
latents: Tensor,
output_height: int,
output_width: int,
out_channels: int,
) -> Tuple[Tensor, Tensor]:
output_height = output_height // self._patch_size[1]
output_width = output_width // self._patch_size[2]
latents = rearrange(
latents,
"b (f h w) (c p q) -> b c f (h p) (w q)",
h=output_height,
w=output_width,
p=self._patch_size[1],
q=self._patch_size[2],
)
return latents
================================================
FILE: ltx_video/models/transformers/transformer3d.py
================================================
# Adapted from: https://github.com/huggingface/diffusers/blob/v0.26.3/src/diffusers/models/transformers/transformer_2d.py
import math
from dataclasses import dataclass
from typing import Any, Dict, List, Optional, Union
import os
import json
import glob
from pathlib import Path
import torch
from diffusers.configuration_utils import ConfigMixin, register_to_config
from diffusers.models.embeddings import PixArtAlphaTextProjection
from diffusers.models.modeling_utils import ModelMixin
from diffusers.models.normalization import AdaLayerNormSingle
from diffusers.utils import BaseOutput, is_torch_version
from diffusers.utils import logging
from torch import nn
from safetensors import safe_open
from ltx_video.models.transformers.attention import BasicTransformerBlock
from ltx_video.utils.skip_layer_strategy import SkipLayerStrategy
from ltx_video.utils.diffusers_config_mapping import (
diffusers_and_ours_config_mapping,
make_hashable_key,
TRANSFORMER_KEYS_RENAME_DICT,
)
logger = logging.get_logger(__name__)
@dataclass
class Transformer3DModelOutput(BaseOutput):
"""
The output of [`Transformer2DModel`].
Args:
sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` or `(batch size, num_vector_embeds - 1, num_latent_pixels)` if [`Transformer2DModel`] is discrete):
The hidden states output conditioned on the `encoder_hidden_states` input. If discrete, returns probability
distributions for the unnoised latent pixels.
"""
sample: torch.FloatTensor
class Transformer3DModel(ModelMixin, ConfigMixin):
_supports_gradient_checkpointing = True
@register_to_config
def __init__(
self,
num_attention_heads: int = 16,
attention_head_dim: int = 88,
in_channels: Optional[int] = None,
out_channels: Optional[int] = None,
num_layers: int = 1,
dropout: float = 0.0,
norm_num_groups: int = 32,
cross_attention_dim: Optional[int] = None,
attention_bias: bool = False,
num_vector_embeds: Optional[int] = None,
activation_fn: str = "geglu",
num_embeds_ada_norm: Optional[int] = None,
use_linear_projection: bool = False,
only_cross_attention: bool = False,
double_self_attention: bool = False,
upcast_attention: bool = False,
adaptive_norm: str = "single_scale_shift", # 'single_scale_shift' or 'single_scale'
standardization_norm: str = "layer_norm", # 'layer_norm' or 'rms_norm'
norm_elementwise_affine: bool = True,
norm_eps: float = 1e-5,
attention_type: str = "default",
caption_channels: int = None,
use_tpu_flash_attention: bool = False, # if True uses the TPU attention offload ('flash attention')
qk_norm: Optional[str] = None,
positional_embedding_type: str = "rope",
positional_embedding_theta: Optional[float] = None,
positional_embedding_max_pos: Optional[List[int]] = None,
timestep_scale_multiplier: Optional[float] = None,
causal_temporal_positioning: bool = False, # For backward compatibility, will be deprecated
):
super().__init__()
self.use_tpu_flash_attention = (
use_tpu_flash_attention # FIXME: push config down to the attention modules
)
self.use_linear_projection = use_linear_projection
self.num_attention_heads = num_attention_heads
self.attention_head_dim = attention_head_dim
inner_dim = num_attention_heads * attention_head_dim
self.inner_dim = inner_dim
self.patchify_proj = nn.Linear(in_channels, inner_dim, bias=True)
self.positional_embedding_type = positional_embedding_type
self.positional_embedding_theta = positional_embedding_theta
self.positional_embedding_max_pos = positional_embedding_max_pos
self.use_rope = self.positional_embedding_type == "rope"
self.timestep_scale_multiplier = timestep_scale_multiplier
if self.positional_embedding_type == "absolute":
raise ValueError("Absolute positional embedding is no longer supported")
elif self.positional_embedding_type == "rope":
if positional_embedding_theta is None:
raise ValueError(
"If `positional_embedding_type` type is rope, `positional_embedding_theta` must also be defined"
)
if positional_embedding_max_pos is None:
raise ValueError(
"If `positional_embedding_type` type is rope, `positional_embedding_max_pos` must also be defined"
)
# 3. Define transformers blocks
self.transformer_blocks = nn.ModuleList(
[
BasicTransformerBlock(
inner_dim,
num_attention_heads,
attention_head_dim,
dropout=dropout,
cross_attention_dim=cross_attention_dim,
activation_fn=activation_fn,
num_embeds_ada_norm=num_embeds_ada_norm,
attention_bias=attention_bias,
only_cross_attention=only_cross_attention,
double_self_attention=double_self_attention,
upcast_attention=upcast_attention,
adaptive_norm=adaptive_norm,
standardization_norm=standardization_norm,
norm_elementwise_affine=norm_elementwise_affine,
norm_eps=norm_eps,
attention_type=attention_type,
use_tpu_flash_attention=use_tpu_flash_attention,
qk_norm=qk_norm,
use_rope=self.use_rope,
)
for d in range(num_layers)
]
)
# 4. Define output layers
self.out_channels = in_channels if out_channels is None else out_channels
self.norm_out = nn.LayerNorm(inner_dim, elementwise_affine=False, eps=1e-6)
self.scale_shift_table = nn.Parameter(
torch.randn(2, inner_dim) / inner_dim**0.5
)
self.proj_out = nn.Linear(inner_dim, self.out_channels)
self.adaln_single = AdaLayerNormSingle(
inner_dim, use_additional_conditions=False
)
if adaptive_norm == "single_scale":
self.adaln_single.linear = nn.Linear(inner_dim, 4 * inner_dim, bias=True)
self.caption_projection = None
if caption_channels is not None:
self.caption_projection = PixArtAlphaTextProjection(
in_features=caption_channels, hidden_size=inner_dim
)
self.gradient_checkpointing = False
def set_use_tpu_flash_attention(self):
r"""
Function sets the flag in this object and propagates down the children. The flag will enforce the usage of TPU
attention kernel.
"""
logger.info("ENABLE TPU FLASH ATTENTION -> TRUE")
self.use_tpu_flash_attention = True
# push config down to the attention modules
for block in self.transformer_blocks:
block.set_use_tpu_flash_attention()
def create_skip_layer_mask(
self,
batch_size: int,
num_conds: int,
ptb_index: int,
skip_block_list: Optional[List[int]] = None,
):
if skip_block_list is None or len(skip_block_list) == 0:
return None
num_layers = len(self.transformer_blocks)
mask = torch.ones(
(num_layers, batch_size * num_conds), device=self.device, dtype=self.dtype
)
for block_idx in skip_block_list:
mask[block_idx, ptb_index::num_conds] = 0
return mask
def _set_gradient_checkpointing(self, module, value=False):
if hasattr(module, "gradient_checkpointing"):
module.gradient_checkpointing = value
def get_fractional_positions(self, indices_grid):
fractional_positions = torch.stack(
[
indices_grid[:, i] / self.positional_embedding_max_pos[i]
for i in range(3)
],
dim=-1,
)
return fractional_positions
def precompute_freqs_cis(self, indices_grid, spacing="exp"):
dtype = torch.float32 # We need full precision in the freqs_cis computation.
dim = self.inner_dim
theta = self.positional_embedding_theta
fractional_positions = self.get_fractional_positions(indices_grid)
start = 1
end = theta
device = fractional_positions.device
if spacing == "exp":
indices = theta ** (
torch.linspace(
math.log(start, theta),
math.log(end, theta),
dim // 6,
device=device,
dtype=dtype,
)
)
indices = indices.to(dtype=dtype)
elif spacing == "exp_2":
indices = 1.0 / theta ** (torch.arange(0, dim, 6, device=device) / dim)
indices = indices.to(dtype=dtype)
elif spacing == "linear":
indices = torch.linspace(start, end, dim // 6, device=device, dtype=dtype)
elif spacing == "sqrt":
indices = torch.linspace(
start**2, end**2, dim // 6, device=device, dtype=dtype
).sqrt()
indices = indices * math.pi / 2
if spacing == "exp_2":
freqs = (
(indices * fractional_positions.unsqueeze(-1))
.transpose(-1, -2)
.flatten(2)
)
else:
freqs = (
(indices * (fractional_positions.unsqueeze(-1) * 2 - 1))
.transpose(-1, -2)
.flatten(2)
)
cos_freq = freqs.cos().repeat_interleave(2, dim=-1)
sin_freq = freqs.sin().repeat_interleave(2, dim=-1)
if dim % 6 != 0:
cos_padding = torch.ones_like(cos_freq[:, :, : dim % 6])
sin_padding = torch.zeros_like(cos_freq[:, :, : dim % 6])
cos_freq = torch.cat([cos_padding, cos_freq], dim=-1)
sin_freq = torch.cat([sin_padding, sin_freq], dim=-1)
return cos_freq.to(self.dtype), sin_freq.to(self.dtype)
def load_state_dict(
self,
state_dict: Dict,
*args,
**kwargs,
):
if any([key.startswith("model.diffusion_model.") for key in state_dict.keys()]):
state_dict = {
key.replace("model.diffusion_model.", ""): value
for key, value in state_dict.items()
if key.startswith("model.diffusion_model.")
}
super().load_state_dict(state_dict, *args, **kwargs)
@classmethod
def from_pretrained(
cls,
pretrained_model_path: Optional[Union[str, os.PathLike]],
*args,
**kwargs,
):
pretrained_model_path = Path(pretrained_model_path)
if pretrained_model_path.is_dir():
config_path = pretrained_model_path / "transformer" / "config.json"
with open(config_path, "r") as f:
config = make_hashable_key(json.load(f))
assert config in diffusers_and_ours_config_mapping, (
"Provided diffusers checkpoint config for transformer is not suppported. "
"We only support diffusers configs found in Lightricks/LTX-Video."
)
config = diffusers_and_ours_config_mapping[config]
state_dict = {}
ckpt_paths = (
pretrained_model_path
/ "transformer"
/ "diffusion_pytorch_model*.safetensors"
)
dict_list = glob.glob(str(ckpt_paths))
for dict_path in dict_list:
part_dict = {}
with safe_open(dict_path, framework="pt", device="cpu") as f:
for k in f.keys():
part_dict[k] = f.get_tensor(k)
state_dict.update(part_dict)
for key in list(state_dict.keys()):
new_key = key
for replace_key, rename_key in TRANSFORMER_KEYS_RENAME_DICT.items():
new_key = new_key.replace(replace_key, rename_key)
state_dict[new_key] = state_dict.pop(key)
with torch.device("meta"):
transformer = cls.from_config(config)
transformer.load_state_dict(state_dict, assign=True, strict=True)
elif pretrained_model_path.is_file() and str(pretrained_model_path).endswith(
".safetensors"
):
comfy_single_file_state_dict = {}
with safe_open(pretrained_model_path, framework="pt", device="cpu") as f:
metadata = f.metadata()
for k in f.keys():
comfy_single_file_state_dict[k] = f.get_tensor(k)
configs = json.loads(metadata["config"])
transformer_config = configs["transformer"]
with torch.device("meta"):
transformer = Transformer3DModel.from_config(transformer_config)
transformer.load_state_dict(comfy_single_file_state_dict, assign=True)
return transformer
def forward(
self,
hidden_states: torch.Tensor,
indices_grid: torch.Tensor,
encoder_hidden_states: Optional[torch.Tensor] = None,
timestep: Optional[torch.LongTensor] = None,
class_labels: Optional[torch.LongTensor] = None,
cross_attention_kwargs: Dict[str, Any] = None,
attention_mask: Optional[torch.Tensor] = None,
encoder_attention_mask: Optional[torch.Tensor] = None,
skip_layer_mask: Optional[torch.Tensor] = None,
skip_layer_strategy: Optional[SkipLayerStrategy] = None,
return_dict: bool = True,
):
"""
The [`Transformer2DModel`] forward method.
Args:
hidden_states (`torch.LongTensor` of shape `(batch size, num latent pixels)` if discrete, `torch.FloatTensor` of shape `(batch size, channel, height, width)` if continuous):
Input `hidden_states`.
indices_grid (`torch.LongTensor` of shape `(batch size, 3, num latent pixels)`):
encoder_hidden_states ( `torch.FloatTensor` of shape `(batch size, sequence len, embed dims)`, *optional*):
Conditional embeddings for cross attention layer. If not given, cross-attention defaults to
self-attention.
timestep ( `torch.LongTensor`, *optional*):
Used to indicate denoising step. Optional timestep to be applied as an embedding in `AdaLayerNorm`.
class_labels ( `torch.LongTensor` of shape `(batch size, num classes)`, *optional*):
Used to indicate class labels conditioning. Optional class labels to be applied as an embedding in
`AdaLayerZeroNorm`.
cross_attention_kwargs ( `Dict[str, Any]`, *optional*):
A kwargs dictionary that if specified is passed along to the `AttentionProcessor` as defined under
`self.processor` in
[diffusers.models.attention_processor](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/attention_processor.py).
attention_mask ( `torch.Tensor`, *optional*):
An attention mask of shape `(batch, key_tokens)` is applied to `encoder_hidden_states`. If `1` the mask
is kept, otherwise if `0` it is discarded. Mask will be converted into a bias, which adds large
negative values to the attention scores corresponding to "discard" tokens.
encoder_attention_mask ( `torch.Tensor`, *optional*):
Cross-attention mask applied to `encoder_hidden_states`. Two formats supported:
* Mask `(batch, sequence_length)` True = keep, False = discard.
* Bias `(batch, 1, sequence_length)` 0 = keep, -10000 = discard.
If `ndim == 2`: will be interpreted as a mask, then converted into a bias consistent with the format
above. This bias will be added to the cross-attention scores.
skip_layer_mask ( `torch.Tensor`, *optional*):
A mask of shape `(num_layers, batch)` that indicates which layers to skip. `0` at position
`layer, batch_idx` indicates that the layer should be skipped for the corresponding batch index.
skip_layer_strategy ( `SkipLayerStrategy`, *optional*, defaults to `None`):
Controls which layers are skipped when calculating a perturbed latent for spatiotemporal guidance.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~models.unets.unet_2d_condition.UNet2DConditionOutput`] instead of a plain
tuple.
Returns:
If `return_dict` is True, an [`~models.transformer_2d.Transformer2DModelOutput`] is returned, otherwise a
`tuple` where the first element is the sample tensor.
"""
# for tpu attention offload 2d token masks are used. No need to transform.
if not self.use_tpu_flash_attention:
# ensure attention_mask is a bias, and give it a singleton query_tokens dimension.
# we may have done this conversion already, e.g. if we came here via UNet2DConditionModel#forward.
# we can tell by counting dims; if ndim == 2: it's a mask rather than a bias.
# expects mask of shape:
# [batch, key_tokens]
# adds singleton query_tokens dimension:
# [batch, 1, key_tokens]
# this helps to broadcast it as a bias over attention scores, which will be in one of the following shapes:
# [batch, heads, query_tokens, key_tokens] (e.g. torch sdp attn)
# [batch * heads, query_tokens, key_tokens] (e.g. xformers or classic attn)
if attention_mask is not None and attention_mask.ndim == 2:
# assume that mask is expressed as:
# (1 = keep, 0 = discard)
# convert mask into a bias that can be added to attention scores:
# (keep = +0, discard = -10000.0)
attention_mask = (1 - attention_mask.to(hidden_states.dtype)) * -10000.0
attention_mask = attention_mask.unsqueeze(1)
# convert encoder_attention_mask to a bias the same way we do for attention_mask
if encoder_attention_mask is not None and encoder_attention_mask.ndim == 2:
encoder_attention_mask = (
1 - encoder_attention_mask.to(hidden_states.dtype)
) * -10000.0
encoder_attention_mask = encoder_attention_mask.unsqueeze(1)
# 1. Input
hidden_states = self.patchify_proj(hidden_states)
if self.timestep_scale_multiplier:
timestep = self.timestep_scale_multiplier * timestep
freqs_cis = self.precompute_freqs_cis(indices_grid)
batch_size = hidden_states.shape[0]
timestep, embedded_timestep = self.adaln_single(
timestep.flatten(),
{"resolution": None, "aspect_ratio": None},
batch_size=batch_size,
hidden_dtype=hidden_states.dtype,
)
# Second dimension is 1 or number of tokens (if timestep_per_token)
timestep = timestep.view(batch_size, -1, timestep.shape[-1])
embedded_timestep = embedded_timestep.view(
batch_size, -1, embedded_timestep.shape[-1]
)
# 2. Blocks
if self.caption_projection is not None:
batch_size = hidden_states.shape[0]
encoder_hidden_states = self.caption_projection(encoder_hidden_states)
encoder_hidden_states = encoder_hidden_states.view(
batch_size, -1, hidden_states.shape[-1]
)
for block_idx, block in enumerate(self.transformer_blocks):
if self.training and self.gradient_checkpointing:
def create_custom_forward(module, return_dict=None):
def custom_forward(*inputs):
if return_dict is not None:
return module(*inputs, return_dict=return_dict)
else:
return module(*inputs)
return custom_forward
ckpt_kwargs: Dict[str, Any] = (
{"use_reentrant": False} if is_torch_version(">=", "1.11.0") else {}
)
hidden_states = torch.utils.checkpoint.checkpoint(
create_custom_forward(block),
hidden_states,
freqs_cis,
attention_mask,
encoder_hidden_states,
encoder_attention_mask,
timestep,
cross_attention_kwargs,
class_labels,
(
skip_layer_mask[block_idx]
if skip_layer_mask is not None
else None
),
skip_layer_strategy,
**ckpt_kwargs,
)
else:
hidden_states = block(
hidden_states,
freqs_cis=freqs_cis,
attention_mask=attention_mask,
encoder_hidden_states=encoder_hidden_states,
encoder_attention_mask=encoder_attention_mask,
timestep=timestep,
cross_attention_kwargs=cross_attention_kwargs,
class_labels=class_labels,
skip_layer_mask=(
skip_layer_mask[block_idx]
if skip_layer_mask is not None
else None
),
skip_layer_strategy=skip_layer_strategy,
)
# 3. Output
scale_shift_values = (
self.scale_shift_table[None, None] + embedded_timestep[:, :, None]
)
shift, scale = scale_shift_values[:, :, 0], scale_shift_values[:, :, 1]
hidden_states = self.norm_out(hidden_states)
# Modulation
hidden_states = hidden_states * (1 + scale) + shift
hidden_states = self.proj_out(hidden_states)
if not return_dict:
return (hidden_states,)
return Transformer3DModelOutput(sample=hidden_states)
================================================
FILE: ltx_video/pipelines/__init__.py
================================================
================================================
FILE: ltx_video/pipelines/crf_compressor.py
================================================
import av
import torch
import io
import numpy as np
def _encode_single_frame(output_file, image_array: np.ndarray, crf):
container = av.open(output_file, "w", format="mp4")
try:
stream = container.add_stream(
"libx264", rate=1, options={"crf": str(crf), "preset": "veryfast"}
)
stream.height = image_array.shape[0]
stream.width = image_array.shape[1]
av_frame = av.VideoFrame.from_ndarray(image_array, format="rgb24").reformat(
format="yuv420p"
)
container.mux(stream.encode(av_frame))
container.mux(stream.encode())
finally:
container.close()
def _decode_single_frame(video_file):
container = av.open(video_file)
try:
stream = next(s for s in container.streams if s.type == "video")
frame = next(container.decode(stream))
finally:
container.close()
return frame.to_ndarray(format="rgb24")
def compress(image: torch.Tensor, crf=29):
if crf == 0:
return image
image_array = (
(image[: (image.shape[0] // 2) * 2, : (image.shape[1] // 2) * 2] * 255.0)
.byte()
.cpu()
.numpy()
)
with io.BytesIO() as output_file:
_encode_single_frame(output_file, image_array, crf)
video_bytes = output_file.getvalue()
with io.BytesIO(video_bytes) as video_file:
image_array = _decode_single_frame(video_file)
tensor = torch.tensor(image_array, dtype=image.dtype, device=image.device) / 255.0
return tensor
================================================
FILE: ltx_video/pipelines/pipeline_ltx_video.py
================================================
# Adapted from: https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/pixart_alpha/pipeline_pixart_alpha.py
import copy
import inspect
import math
import re
from contextlib import nullcontext
from dataclasses import dataclass
from typing import Any, Callable, Dict, List, Optional, Tuple, Union
import torch
import torch.nn.functional as F
from diffusers.image_processor import VaeImageProcessor
from diffusers.models import AutoencoderKL
from diffusers.pipelines.pipeline_utils import DiffusionPipeline, ImagePipelineOutput
from diffusers.schedulers import DPMSolverMultistepScheduler
from diffusers.utils import deprecate, logging
from diffusers.utils.torch_utils import randn_tensor
from einops import rearrange
from transformers import (
T5EncoderModel,
T5Tokenizer,
AutoModelForCausalLM,
AutoProcessor,
AutoTokenizer,
)
from ltx_video.models.autoencoders.causal_video_autoencoder import (
CausalVideoAutoencoder,
)
from ltx_video.models.autoencoders.vae_encode import (
get_vae_size_scale_factor,
latent_to_pixel_coords,
vae_decode,
vae_encode,
)
from ltx_video.models.transformers.symmetric_patchifier import Patchifier
from ltx_video.models.transformers.transformer3d import Transformer3DModel
from ltx_video.schedulers.rf import TimestepShifter
from ltx_video.utils.skip_layer_strategy import SkipLayerStrategy
from ltx_video.utils.prompt_enhance_utils import generate_cinematic_prompt
from ltx_video.models.autoencoders.latent_upsampler import LatentUpsampler
from ltx_video.models.autoencoders.vae_encode import (
un_normalize_latents,
normalize_latents,
)
logger = logging.get_logger(__name__) # pylint: disable=invalid-name
ASPECT_RATIO_1024_BIN = {
"0.25": [512.0, 2048.0],
"0.28": [512.0, 1856.0],
"0.32": [576.0, 1792.0],
"0.33": [576.0, 1728.0],
"0.35": [576.0, 1664.0],
"0.4": [640.0, 1600.0],
"0.42": [640.0, 1536.0],
"0.48": [704.0, 1472.0],
"0.5": [704.0, 1408.0],
"0.52": [704.0, 1344.0],
"0.57": [768.0, 1344.0],
"0.6": [768.0, 1280.0],
"0.68": [832.0, 1216.0],
"0.72": [832.0, 1152.0],
"0.78": [896.0, 1152.0],
"0.82": [896.0, 1088.0],
"0.88": [960.0, 1088.0],
"0.94": [960.0, 1024.0],
"1.0": [1024.0, 1024.0],
"1.07": [1024.0, 960.0],
"1.13": [1088.0, 960.0],
"1.21": [1088.0, 896.0],
"1.29": [1152.0, 896.0],
"1.38": [1152.0, 832.0],
"1.46": [1216.0, 832.0],
"1.67": [1280.0, 768.0],
"1.75": [1344.0, 768.0],
"2.0": [1408.0, 704.0],
"2.09": [1472.0, 704.0],
"2.4": [1536.0, 640.0],
"2.5": [1600.0, 640.0],
"3.0": [1728.0, 576.0],
"4.0": [2048.0, 512.0],
}
ASPECT_RATIO_512_BIN = {
"0.25": [256.0, 1024.0],
"0.28": [256.0, 928.0],
"0.32": [288.0, 896.0],
"0.33": [288.0, 864.0],
"0.35": [288.0, 832.0],
"0.4": [320.0, 800.0],
"0.42": [320.0, 768.0],
"0.48": [352.0, 736.0],
"0.5": [352.0, 704.0],
"0.52": [352.0, 672.0],
"0.57": [384.0, 672.0],
"0.6": [384.0, 640.0],
"0.68": [416.0, 608.0],
"0.72": [416.0, 576.0],
"0.78": [448.0, 576.0],
"0.82": [448.0, 544.0],
"0.88": [480.0, 544.0],
"0.94": [480.0, 512.0],
"1.0": [512.0, 512.0],
"1.07": [512.0, 480.0],
"1.13": [544.0, 480.0],
"1.21": [544.0, 448.0],
"1.29": [576.0, 448.0],
"1.38": [576.0, 416.0],
"1.46": [608.0, 416.0],
"1.67": [640.0, 384.0],
"1.75": [672.0, 384.0],
"2.0": [704.0, 352.0],
"2.09": [736.0, 352.0],
"2.4": [768.0, 320.0],
"2.5": [800.0, 320.0],
"3.0": [864.0, 288.0],
"4.0": [1024.0, 256.0],
}
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.retrieve_timesteps
def retrieve_timesteps(
scheduler,
num_inference_steps: Optional[int] = None,
device: Optional[Union[str, torch.device]] = None,
timesteps: Optional[List[int]] = None,
skip_initial_inference_steps: int = 0,
skip_final_inference_steps: int = 0,
**kwargs,
):
"""
Calls the scheduler's `set_timesteps` method and retrieves timesteps from the scheduler after the call. Handles
custom timesteps. Any kwargs will be supplied to `scheduler.set_timesteps`.
Args:
scheduler (`SchedulerMixin`):
The scheduler to get timesteps from.
num_inference_steps (`int`):
The number of diffusion steps used when generating samples with a pre-trained model. If used,
`timesteps` must be `None`.
device (`str` or `torch.device`, *optional*):
The device to which the timesteps should be moved to. If `None`, the timesteps are not moved.
timesteps (`List[int]`, *optional*):
Custom timesteps used to support arbitrary spacing between timesteps. If `None`, then the default
timestep spacing strategy of the scheduler is used. If `timesteps` is passed, `num_inference_steps`
must be `None`.
max_timestep ('float', *optional*, defaults to 1.0):
The initial noising level for image-to-image/video-to-video. The list if timestamps will be
truncated to start with a timestamp greater or equal to this.
Returns:
`Tuple[torch.Tensor, int]`: A tuple where the first element is the timestep schedule from the scheduler and the
second element is the number of inference steps.
"""
if timesteps is not None:
accepts_timesteps = "timesteps" in set(
inspect.signature(scheduler.set_timesteps).parameters.keys()
)
if not accepts_timesteps:
raise ValueError(
f"The current scheduler class {scheduler.__class__}'s `set_timesteps` does not support custom"
f" timestep schedules. Please check whether you are using the correct scheduler."
)
scheduler.set_timesteps(timesteps=timesteps, device=device, **kwargs)
timesteps = scheduler.timesteps
num_inference_steps = len(timesteps)
else:
scheduler.set_timesteps(num_inference_steps, device=device, **kwargs)
timesteps = scheduler.timesteps
if (
skip_initial_inference_steps < 0
or skip_final_inference_steps < 0
or skip_initial_inference_steps + skip_final_inference_steps
>= num_inference_steps
):
raise ValueError(
"invalid skip inference step values: must be non-negative and the sum of skip_initial_inference_steps and skip_final_inference_steps must be less than the number of inference steps"
)
timesteps = timesteps[
skip_initial_inference_steps : len(timesteps) - skip_final_inference_steps
]
scheduler.set_timesteps(timesteps=timesteps, device=device, **kwargs)
num_inference_steps = len(timesteps)
return timesteps, num_inference_steps
@dataclass
class ConditioningItem:
"""
Defines a single frame-conditioning item - a single frame or a sequence of frames.
Attributes:
media_item (torch.Tensor): shape=(b, 3, f, h, w). The media item to condition on.
media_frame_number (int): The start-frame number of the media item in the generated video.
conditioning_strength (float): The strength of the conditioning (1.0 = full conditioning).
media_x (Optional[int]): Optional left x coordinate of the media item in the generated frame.
media_y (Optional[int]): Optional top y coordinate of the media item in the generated frame.
"""
media_item: torch.Tensor
media_frame_number: int
conditioning_strength: float
media_x: Optional[int] = None
media_y: Optional[int] = None
class LTXVideoPipeline(DiffusionPipeline):
r"""
Pipeline for text-to-image generation using LTX-Video.
This model inherits from [`DiffusionPipeline`]. Check the superclass documentation for the generic methods the
library implements for all the pipelines (such as downloading or saving, running on a particular device, etc.)
Args:
vae ([`AutoencoderKL`]):
Variational Auto-Encoder (VAE) Model to encode and decode images to and from latent representations.
text_encoder ([`T5EncoderModel`]):
Frozen text-encoder. This uses
[T5](https://huggingface.co/docs/transformers/model_doc/t5#transformers.T5EncoderModel), specifically the
[t5-v1_1-xxl](https://huggingface.co/PixArt-alpha/PixArt-alpha/tree/main/t5-v1_1-xxl) variant.
tokenizer (`T5Tokenizer`):
Tokenizer of class
[T5Tokenizer](https://huggingface.co/docs/transformers/model_doc/t5#transformers.T5Tokenizer).
transformer ([`Transformer2DModel`]):
A text conditioned `Transformer2DModel` to denoise the encoded image latents.
scheduler ([`SchedulerMixin`]):
A scheduler to be used in combination with `transformer` to denoise the encoded image latents.
"""
bad_punct_regex = re.compile(
r"["
+ "#®•©™&@·º½¾¿¡§~"
+ r"\)"
+ r"\("
+ r"\]"
+ r"\["
+ r"\}"
+ r"\{"
+ r"\|"
+ "\\"
+ r"\/"
+ r"\*"
+ r"]{1,}"
) # noqa
_optional_components = [
"tokenizer",
"text_encoder",
"prompt_enhancer_image_caption_model",
"prompt_enhancer_image_caption_processor",
"prompt_enhancer_llm_model",
"prompt_enhancer_llm_tokenizer",
]
model_cpu_offload_seq = "prompt_enhancer_image_caption_model->prompt_enhancer_llm_model->text_encoder->transformer->vae"
def __init__(
self,
tokenizer: T5Tokenizer,
text_encoder: T5EncoderModel,
vae: AutoencoderKL,
transformer: Transformer3DModel,
scheduler: DPMSolverMultistepScheduler,
patchifier: Patchifier,
prompt_enhancer_image_caption_model: AutoModelForCausalLM,
prompt_enhancer_image_caption_processor: AutoProcessor,
prompt_enhancer_llm_model: AutoModelForCausalLM,
prompt_enhancer_llm_tokenizer: AutoTokenizer,
allowed_inference_steps: Optional[List[float]] = None,
):
super().__init__()
self.register_modules(
tokenizer=tokenizer,
text_encoder=text_encoder,
vae=vae,
transformer=transformer,
scheduler=scheduler,
patchifier=patchifier,
prompt_enhancer_image_caption_model=prompt_enhancer_image_caption_model,
prompt_enhancer_image_caption_processor=prompt_enhancer_image_caption_processor,
prompt_enhancer_llm_model=prompt_enhancer_llm_model,
prompt_enhancer_llm_tokenizer=prompt_enhancer_llm_tokenizer,
)
self.video_scale_factor, self.vae_scale_factor, _ = get_vae_size_scale_factor(
self.vae
)
self.image_processor = VaeImageProcessor(vae_scale_factor=self.vae_scale_factor)
self.allowed_inference_steps = allowed_inference_steps
def mask_text_embeddings(self, emb, mask):
if emb.shape[0] == 1:
keep_index = mask.sum().item()
return emb[:, :, :keep_index, :], keep_index
else:
masked_feature = emb * mask[:, None, :, None]
return masked_feature, emb.shape[2]
# Adapted from diffusers.pipelines.deepfloyd_if.pipeline_if.encode_prompt
def encode_prompt(
self,
prompt: Union[str, List[str]],
do_classifier_free_guidance: bool = True,
negative_prompt: str = "",
num_images_per_prompt: int = 1,
device: Optional[torch.device] = None,
prompt_embeds: Optional[torch.FloatTensor] = None,
negative_prompt_embeds: Optional[torch.FloatTensor] = None,
prompt_attention_mask: Optional[torch.FloatTensor] = None,
negative_prompt_attention_mask: Optional[torch.FloatTensor] = None,
text_encoder_max_tokens: int = 256,
**kwargs,
):
r"""
Encodes the prompt into text encoder hidden states.
Args:
prompt (`str` or `List[str]`, *optional*):
prompt to be encoded
negative_prompt (`str` or `List[str]`, *optional*):
The prompt not to guide the image generation. If not defined, one has to pass `negative_prompt_embeds`
instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is less than `1`). For
This should be "".
do_classifier_free_guidance (`bool`, *optional*, defaults to `True`):
whether to use classifier free guidance or not
num_images_per_prompt (`int`, *optional*, defaults to 1):
number of images that should be generated per prompt
device: (`torch.device`, *optional*):
torch device to place the resulting embeddings on
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
provided, text embeddings will be generated from `prompt` input argument.
negative_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated negative text embeddings.
"""
if "mask_feature" in kwargs:
deprecation_message = "The use of `mask_feature` is deprecated. It is no longer used in any computation and that doesn't affect the end results. It will be removed in a future version."
deprecate("mask_feature", "1.0.0", deprecation_message, standard_warn=False)
if device is None:
device = self._execution_device
if prompt is not None and isinstance(prompt, str):
batch_size = 1
elif prompt is not None and isinstance(prompt, list):
batch_size = len(prompt)
else:
batch_size = prompt_embeds.shape[0]
# See Section 3.1. of the paper.
max_length = (
text_encoder_max_tokens # TPU supports only lengths multiple of 128
)
if prompt_embeds is None:
assert (
self.text_encoder is not None
), "You should provide either prompt_embeds or self.text_encoder should not be None,"
text_enc_device = next(self.text_encoder.parameters()).device
prompt = self._text_preprocessing(prompt)
text_inputs = self.tokenizer(
prompt,
padding="max_length",
max_length=max_length,
truncation=True,
add_special_tokens=True,
return_tensors="pt",
)
text_input_ids = text_inputs.input_ids
untruncated_ids = self.tokenizer(
prompt, padding="longest", return_tensors="pt"
).input_ids
if untruncated_ids.shape[-1] >= text_input_ids.shape[
-1
] and not torch.equal(text_input_ids, untruncated_ids):
removed_text = self.tokenizer.batch_decode(
untruncated_ids[:, max_length - 1 : -1]
)
logger.warning(
"The following part of your input was truncated because CLIP can only handle sequences up to"
f" {max_length} tokens: {removed_text}"
)
prompt_attention_mask = text_inputs.attention_mask
prompt_attention_mask = prompt_attention_mask.to(text_enc_device)
prompt_attention_mask = prompt_attention_mask.to(device)
prompt_embeds = self.text_encoder(
text_input_ids.to(text_enc_device), attention_mask=prompt_attention_mask
)
prompt_embeds = prompt_embeds[0]
if self.text_encoder is not None:
dtype = self.text_encoder.dtype
elif self.transformer is not None:
dtype = self.transformer.dtype
else:
dtype = None
prompt_embeds = prompt_embeds.to(dtype=dtype, device=device)
bs_embed, seq_len, _ = prompt_embeds.shape
# duplicate text embeddings and attention mask for each generation per prompt, using mps friendly method
prompt_embeds = prompt_embeds.repeat(1, num_images_per_prompt, 1)
prompt_embeds = prompt_embeds.view(
bs_embed * num_images_per_prompt, seq_len, -1
)
prompt_attention_mask = prompt_attention_mask.repeat(1, num_images_per_prompt)
prompt_attention_mask = prompt_attention_mask.view(
bs_embed * num_images_per_prompt, -1
)
# get unconditional embeddings for classifier free guidance
if do_classifier_free_guidance and negative_prompt_embeds is None:
uncond_tokens = self._text_preprocessing(negative_prompt)
uncond_tokens = uncond_tokens * batch_size
max_length = prompt_embeds.shape[1]
uncond_input = self.tokenizer(
uncond_tokens,
padding="max_length",
max_length=max_length,
truncation=True,
return_attention_mask=True,
add_special_tokens=True,
return_tensors="pt",
)
negative_prompt_attention_mask = uncond_input.attention_mask
negative_prompt_attention_mask = negative_prompt_attention_mask.to(
text_enc_device
)
negative_prompt_embeds = self.text_encoder(
uncond_input.input_ids.to(text_enc_device),
attention_mask=negative_prompt_attention_mask,
)
negative_prompt_embeds = negative_prompt_embeds[0]
if do_classifier_free_guidance:
# duplicate unconditional embeddings for each generation per prompt, using mps friendly method
seq_len = negative_prompt_embeds.shape[1]
negative_prompt_embeds = negative_prompt_embeds.to(
dtype=dtype, device=device
)
negative_prompt_embeds = negative_prompt_embeds.repeat(
1, num_images_per_prompt, 1
)
negative_prompt_embeds = negative_prompt_embeds.view(
batch_size * num_images_per_prompt, seq_len, -1
)
negative_prompt_attention_mask = negative_prompt_attention_mask.repeat(
1, num_images_per_prompt
)
negative_prompt_attention_mask = negative_prompt_attention_mask.view(
bs_embed * num_images_per_prompt, -1
)
else:
negative_prompt_embeds = None
negative_prompt_attention_mask = None
return (
prompt_embeds,
prompt_attention_mask,
negative_prompt_embeds,
negative_prompt_attention_mask,
)
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_extra_step_kwargs
def prepare_extra_step_kwargs(self, generator, eta):
# prepare extra kwargs for the scheduler step, since not all schedulers have the same signature
# eta (η) is only used with the DDIMScheduler, it will be ignored for other schedulers.
# eta corresponds to η in DDIM paper: https://arxiv.org/abs/2010.02502
# and should be between [0, 1]
accepts_eta = "eta" in set(
inspect.signature(self.scheduler.step).parameters.keys()
)
extra_step_kwargs = {}
if accepts_eta:
extra_step_kwargs["eta"] = eta
# check if the scheduler accepts generator
accepts_generator = "generator" in set(
inspect.signature(self.scheduler.step).parameters.keys()
)
if accepts_generator:
extra_step_kwargs["generator"] = generator
return extra_step_kwargs
def check_inputs(
self,
prompt,
height,
width,
negative_prompt,
prompt_embeds=None,
negative_prompt_embeds=None,
prompt_attention_mask=None,
negative_prompt_attention_mask=None,
enhance_prompt=False,
):
if height % 8 != 0 or width % 8 != 0:
raise ValueError(
f"`height` and `width` have to be divisible by 8 but are {height} and {width}."
)
if prompt is not None and prompt_embeds is not None:
raise ValueError(
f"Cannot forward both `prompt`: {prompt} and `prompt_embeds`: {prompt_embeds}. Please make sure to"
" only forward one of the two."
)
elif prompt is None and prompt_embeds is None:
raise ValueError(
"Provide either `prompt` or `prompt_embeds`. Cannot leave both `prompt` and `prompt_embeds` undefined."
)
elif prompt is not None and (
not isinstance(prompt, str) and not isinstance(prompt, list)
):
raise ValueError(
f"`prompt` has to be of type `str` or `list` but is {type(prompt)}"
)
if prompt is not None and negative_prompt_embeds is not None:
raise ValueError(
f"Cannot forward both `prompt`: {prompt} and `negative_prompt_embeds`:"
f" {negative_prompt_embeds}. Please make sure to only forward one of the two."
)
if negative_prompt is not None and negative_prompt_embeds is not None:
raise ValueError(
f"Cannot forward both `negative_prompt`: {negative_prompt} and `negative_prompt_embeds`:"
f" {negative_prompt_embeds}. Please make sure to only forward one of the two."
)
if prompt_embeds is not None and prompt_attention_mask is None:
raise ValueError(
"Must provide `prompt_attention_mask` when specifying `prompt_embeds`."
)
if (
negative_prompt_embeds is not None
and negative_prompt_attention_mask is None
):
raise ValueError(
"Must provide `negative_prompt_attention_mask` when specifying `negative_prompt_embeds`."
)
if prompt_embeds is not None and negative_prompt_embeds is not None:
if prompt_embeds.shape != negative_prompt_embeds.shape:
raise ValueError(
"`prompt_embeds` and `negative_prompt_embeds` must have the same shape when passed directly, but"
f" got: `prompt_embeds` {prompt_embeds.shape} != `negative_prompt_embeds`"
f" {negative_prompt_embeds.shape}."
)
if prompt_attention_mask.shape != negative_prompt_attention_mask.shape:
raise ValueError(
"`prompt_attention_mask` and `negative_prompt_attention_mask` must have the same shape when passed directly, but"
f" got: `prompt_attention_mask` {prompt_attention_mask.shape} != `negative_prompt_attention_mask`"
f" {negative_prompt_attention_mask.shape}."
)
if enhance_prompt:
assert (
self.prompt_enhancer_image_caption_model is not None
), "Image caption model must be initialized if enhance_prompt is True"
assert (
self.prompt_enhancer_image_caption_processor is not None
), "Image caption processor must be initialized if enhance_prompt is True"
assert (
self.prompt_enhancer_llm_model is not None
), "Text prompt enhancer model must be initialized if enhance_prompt is True"
assert (
self.prompt_enhancer_llm_tokenizer is not None
), "Text prompt enhancer tokenizer must be initialized if enhance_prompt is True"
def _text_preprocessing(self, text):
if not isinstance(text, (tuple, list)):
text = [text]
def process(text: str):
text = text.strip()
return text
return [process(t) for t in text]
@staticmethod
def add_noise_to_image_conditioning_latents(
t: float,
init_latents: torch.Tensor,
latents: torch.Tensor,
noise_scale: float,
conditioning_mask: torch.Tensor,
generator,
eps=1e-6,
):
"""
Add timestep-dependent noise to the hard-conditioning latents.
This helps with motion continuity, especially when conditioned on a single frame.
"""
noise = randn_tensor(
latents.shape,
generator=generator,
device=latents.device,
dtype=latents.dtype,
)
# Add noise only to hard-conditioning latents (conditioning_mask = 1.0)
need_to_noise = (conditioning_mask > 1.0 - eps).unsqueeze(-1)
noised_latents = init_latents + noise_scale * noise * (t**2)
latents = torch.where(need_to_noise, noised_latents, latents)
return latents
# Copied from diffusers.pipelines.stable_diffusion.pipeline_stable_diffusion.StableDiffusionPipeline.prepare_latents
def prepare_latents(
self,
latents: torch.Tensor | None,
media_items: torch.Tensor | None,
timestep: float,
latent_shape: torch.Size | Tuple[Any, ...],
dtype: torch.dtype,
device: torch.device,
generator: torch.Generator | List[torch.Generator],
vae_per_channel_normalize: bool = True,
):
"""
Prepare the initial latent tensor to be denoised.
The latents are either pure noise or a noised version of the encoded media items.
Args:
latents (`torch.FloatTensor` or `None`):
The latents to use (provided by the user) or `None` to create new latents.
media_items (`torch.FloatTensor` or `None`):
An image or video to be updated using img2img or vid2vid. The media item is encoded and noised.
timestep (`float`):
The timestep to noise the encoded media_items to.
latent_shape (`torch.Size`):
The target latent shape.
dtype (`torch.dtype`):
The target dtype.
device (`torch.device`):
The target device.
generator (`torch.Generator` or `List[torch.Generator]`):
Generator(s) to be used for the noising process.
vae_per_channel_normalize ('bool'):
When encoding the media_items, whether to normalize the latents per-channel.
Returns:
`torch.FloatTensor`: The latents to be used for the denoising process. This is a tensor of shape
(batch_size, num_channels, height, width).
"""
if isinstance(generator, list) and len(generator) != latent_shape[0]:
raise ValueError(
f"You have passed a list of generators of length {len(generator)}, but requested an effective batch"
f" size of {latent_shape[0]}. Make sure the batch size matches the length of the generators."
)
# Initialize the latents with the given latents or encoded media item, if provided
assert (
latents is None or media_items is None
), "Cannot provide both latents and media_items. Please provide only one of the two."
assert (
latents is None and media_items is None or timestep < 1.0
), "Input media_item or latents are provided, but they will be replaced with noise."
if media_items is not None:
latents = vae_encode(
media_items.to(dtype=self.vae.dtype, device=self.vae.device),
self.vae,
vae_per_channel_normalize=vae_per_channel_normalize,
)
if latents is not None:
assert (
latents.shape == latent_shape
), f"Latents have to be of shape {latent_shape} but are {latents.shape}."
latents = latents.to(device=device, dtype=dtype)
# For backward compatibility, generate in the "patchified" shape and rearrange
b, c, f, h, w = latent_shape
noise = randn_tensor(
(b, f * h * w, c), generator=generator, device=device, dtype=dtype
)
noise = rearrange(noise, "b (f h w) c -> b c f h w", f=f, h=h, w=w)
# scale the initial noise by the standard deviation required by the scheduler
noise = noise * self.scheduler.init_noise_sigma
if latents is None:
latents = noise
else:
# Noise the latents to the required (first) timestep
latents = timestep * noise + (1 - timestep) * latents
return latents
@staticmethod
def classify_height_width_bin(
height: int, width: int, ratios: dict
) -> Tuple[int, int]:
"""Returns binned height and width."""
ar = float(height / width)
closest_ratio = min(ratios.keys(), key=lambda ratio: abs(float(ratio) - ar))
default_hw = ratios[closest_ratio]
return int(default_hw[0]), int(default_hw[1])
@staticmethod
def resize_and_crop_tensor(
samples: torch.Tensor, new_width: int, new_height: int
) -> torch.Tensor:
n_frames, orig_height, orig_width = samples.shape[-3:]
# Check if resizing is needed
if orig_height != new_height or orig_width != new_width:
ratio = max(new_height / orig_height, new_width / orig_width)
resized_width = int(orig_width * ratio)
resized_height = int(orig_height * ratio)
# Resize
samples = LTXVideoPipeline.resize_tensor(
samples, resized_height, resized_width
)
# Center Crop
start_x = (resized_width - new_width) // 2
end_x = start_x + new_width
start_y = (resized_height - new_height) // 2
end_y = start_y + new_height
samples = samples[..., start_y:end_y, start_x:end_x]
return samples
@staticmethod
def resize_tensor(media_items, height, width):
n_frames = media_items.shape[2]
if media_items.shape[-2:] != (height, width):
media_items = rearrange(media_items, "b c n h w -> (b n) c h w")
media_items = F.interpolate(
media_items,
size=(height, width),
mode="bilinear",
align_corners=False,
)
media_items = rearrange(media_items, "(b n) c h w -> b c n h w", n=n_frames)
return media_items
@torch.no_grad()
def __call__(
self,
height: int,
width: int,
num_frames: int,
frame_rate: float,
prompt: Union[str, List[str]] = None,
negative_prompt: str = "",
num_inference_steps: int = 20,
skip_initial_inference_steps: int = 0,
skip_final_inference_steps: int = 0,
timesteps: List[int] = None,
guidance_scale: Union[float, List[float]] = 4.5,
cfg_star_rescale: bool = False,
skip_layer_strategy: Optional[SkipLayerStrategy] = None,
skip_block_list: Optional[Union[List[List[int]], List[int]]] = None,
stg_scale: Union[float, List[float]] = 1.0,
rescaling_scale: Union[float, List[float]] = 0.7,
guidance_timesteps: Optional[List[int]] = None,
num_images_per_prompt: Optional[int] = 1,
eta: float = 0.0,
generator: Optional[Union[torch.Generator, List[torch.Generator]]] = None,
latents: Optional[torch.FloatTensor] = None,
prompt_embeds: Optional[torch.FloatTensor] = None,
prompt_attention_mask: Optional[torch.FloatTensor] = None,
negative_prompt_embeds: Optional[torch.FloatTensor] = None,
negative_prompt_attention_mask: Optional[torch.FloatTensor] = None,
output_type: Optional[str] = "pil",
return_dict: bool = True,
callback_on_step_end: Optional[Callable[[int, int, Dict], None]] = None,
conditioning_items: Optional[List[ConditioningItem]] = None,
decode_timestep: Union[List[float], float] = 0.0,
decode_noise_scale: Optional[List[float]] = None,
mixed_precision: bool = False,
offload_to_cpu: bool = False,
enhance_prompt: bool = False,
text_encoder_max_tokens: int = 256,
stochastic_sampling: bool = False,
media_items: Optional[torch.Tensor] = None,
tone_map_compression_ratio: float = 0.0,
**kwargs,
) -> Union[ImagePipelineOutput, Tuple]:
"""
Function invoked when calling the pipeline for generation.
Args:
prompt (`str` or `List[str]`, *optional*):
The prompt or prompts to guide the image generation. If not defined, one has to pass `prompt_embeds`.
instead.
negative_prompt (`str` or `List[str]`, *optional*):
The prompt or prompts not to guide the image generation. If not defined, one has to pass
`negative_prompt_embeds` instead. Ignored when not using guidance (i.e., ignored if `guidance_scale` is
less than `1`).
num_inference_steps (`int`, *optional*, defaults to 100):
The number of denoising steps. More denoising steps usually lead to a higher quality image at the
expense of slower inference. If `timesteps` is provided, this parameter is ignored.
skip_initial_inference_steps (`int`, *optional*, defaults to 0):
The number of initial timesteps to skip. After calculating the timesteps, this number of timesteps will
be removed from the beginning of the timesteps list. Meaning the highest-timesteps values will not run.
skip_final_inference_steps (`int`, *optional*, defaults to 0):
The number of final timesteps to skip. After calculating the timesteps, this number of timesteps will
be removed from the end of the timesteps list. Meaning the lowest-timesteps values will not run.
timesteps (`List[int]`, *optional*):
Custom timesteps to use for the denoising process. If not defined, equal spaced `num_inference_steps`
timesteps are used. Must be in descending order.
guidance_scale (`float`, *optional*, defaults to 4.5):
Guidance scale as defined in [Classifier-Free Diffusion Guidance](https://arxiv.org/abs/2207.12598).
`guidance_scale` is defined as `w` of equation 2. of [Imagen
Paper](https://arxiv.org/pdf/2205.11487.pdf). Guidance scale is enabled by setting `guidance_scale >
1`. Higher guidance scale encourages to generate images that are closely linked to the text `prompt`,
usually at the expense of lower image quality.
cfg_star_rescale (`bool`, *optional*, defaults to `False`):
If set to `True`, applies the CFG star rescale. Scales the negative prediction according to dot
product between positive and negative.
num_images_per_prompt (`int`, *optional*, defaults to 1):
The number of images to generate per prompt.
height (`int`, *optional*, defaults to self.unet.config.sample_size):
The height in pixels of the generated image.
width (`int`, *optional*, defaults to self.unet.config.sample_size):
The width in pixels of the generated image.
eta (`float`, *optional*, defaults to 0.0):
Corresponds to parameter eta (η) in the DDIM paper: https://arxiv.org/abs/2010.02502. Only applies to
[`schedulers.DDIMScheduler`], will be ignored for others.
generator (`torch.Generator` or `List[torch.Generator]`, *optional*):
One or a list of [torch generator(s)](https://pytorch.org/docs/stable/generated/torch.Generator.html)
to make generation deterministic.
latents (`torch.FloatTensor`, *optional*):
Pre-generated noisy latents, sampled from a Gaussian distribution, to be used as inputs for image
generation. Can be used to tweak the same generation with different prompts. If not provided, a latents
tensor will ge generated by sampling using the supplied random `generator`.
prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated text embeddings. Can be used to easily tweak text inputs, *e.g.* prompt weighting. If not
provided, text embeddings will be generated from `prompt` input argument.
prompt_attention_mask (`torch.FloatTensor`, *optional*): Pre-generated attention mask for text embeddings.
negative_prompt_embeds (`torch.FloatTensor`, *optional*):
Pre-generated negative text embeddings. This negative prompt should be "". If not
provided, negative_prompt_embeds will be generated from `negative_prompt` input argument.
negative_prompt_attention_mask (`torch.FloatTensor`, *optional*):
Pre-generated attention mask for negative text embeddings.
output_type (`str`, *optional*, defaults to `"pil"`):
The output format of the generate image. Choose between
[PIL](https://pillow.readthedocs.io/en/stable/): `PIL.Image.Image` or `np.array`.
return_dict (`bool`, *optional*, defaults to `True`):
Whether to return a [`~pipelines.stable_diffusion.IFPipelineOutput`] instead of a plain tuple.
callback_on_step_end (`Callable`, *optional*):
A function that calls at the end of each denoising steps during the inference. The function is called
with the following arguments: `callback_on_step_end(self: DiffusionPipeline, step: int, timestep: int,
callback_kwargs: Dict)`. `callback_kwargs` will include a list of all tensors as specified by
`callback_on_step_end_tensor_inputs`.
use_resolution_binning (`bool` defaults to `True`):
If set to `True`, the requested height and width are first mapped to the closest resolutions using
`ASPECT_RATIO_1024_BIN`. After the produced latents are decoded into images, they are resized back to
the requested resolution. Useful for generating non-square images.
enhance_prompt (`bool`, *optional*, defaults to `False`):
If set to `True`, the prompt is enhanced using a LLM model.
text_encoder_max_tokens (`int`, *optional*, defaults to `256`):
The maximum number of tokens to use for the text encoder.
stochastic_sampling (`bool`, *optional*, defaults to `False`):
If set to `True`, the sampling is stochastic. If set to `False`, the sampling is deterministic.
media_items ('torch.Tensor', *optional*):
The input media item used for image-to-image / video-to-video.
tone_map_compression_ratio: compression ratio for tone mapping, defaults to 0.0.
If set to 0.0, no tone mapping is applied. If set to 1.0 - full compression is applied.
Examples:
Returns:
[`~pipelines.ImagePipelineOutput`] or `tuple`:
If `return_dict` is `True`, [`~pipelines.ImagePipelineOutput`] is returned, otherwise a `tuple` is
returned where the first element is a list with the generated images
"""
if "mask_feature" in kwargs:
deprecation_message = "The use of `mask_feature` is deprecated. It is no longer used in any computation and that doesn't affect the end results. It will be removed in a future version."
deprecate("mask_feature", "1.0.0", deprecation_message, standard_warn=False)
is_video = kwargs.get("is_video", False)
self.check_inputs(
prompt,
height,
width,
negative_prompt,
prompt_embeds,
negative_prompt_embeds,
prompt_attention_mask,
negative_prompt_attention_mask,
)
# 2. Default height and width to transformer
if prompt is not None and isinstance(prompt, str):
batch_size = 1
elif prompt is not None and isinstance(prompt, list):
batch_size = len(prompt)
else:
batch_size = prompt_embeds.shape[0]
device = self._execution_device
self.video_scale_factor = self.video_scale_factor if is_video else 1
vae_per_channel_normalize = kwargs.get("vae_per_channel_normalize", True)
image_cond_noise_scale = kwargs.get("image_cond_noise_scale", 0.0)
latent_height = height // self.vae_scale_factor
latent_width = width // self.vae_scale_factor
latent_num_frames = num_frames // self.video_scale_factor
if isinstance(self.vae, CausalVideoAutoencoder) and is_video:
latent_num_frames += 1
latent_shape = (
batch_size * num_images_per_prompt,
self.transformer.config.in_channels,
latent_num_frames,
latent_height,
latent_width,
)
# Prepare the list of denoising time-steps
retrieve_timesteps_kwargs = {}
if isinstance(self.scheduler, TimestepShifter):
retrieve_timesteps_kwargs["samples_shape"] = latent_shape
assert (
skip_initial_inference_steps == 0
or latents is not None
or media_items is not None
), (
f"skip_initial_inference_steps ({skip_initial_inference_steps}) is used for image-to-image/video-to-video - "
"media_item or latents should be provided."
)
timesteps, num_inference_steps = retrieve_timesteps(
self.scheduler,
num_inference_steps,
device,
timesteps,
skip_initial_inference_steps=skip_initial_inference_steps,
skip_final_inference_steps=skip_final_inference_steps,
**retrieve_timesteps_kwargs,
)
if self.allowed_inference_steps is not None:
for timestep in [round(x, 4) for x in timesteps.tolist()]:
assert (
timestep in self.allowed_inference_steps
), f"Invalid inference timestep {timestep}. Allowed timesteps are {self.allowed_inference_steps}."
if guidance_timesteps:
guidance_mapping = []
for timestep in timesteps:
indices = [
i for i, val in enumerate(guidance_timesteps) if val <= timestep
]
# assert len(indices) > 0, f"No guidance timestep found for {timestep}"
guidance_mapping.append(
indices[0] if len(indices) > 0 else (len(guidance_timesteps) - 1)
)
# here `guidance_scale` is defined analog to the guidance weight `w` of equation (2)
# of the Imagen paper: https://arxiv.org/pdf/2205.11487.pdf . `guidance_scale = 1`
# corresponds to doing no classifier free guidance.
if not isinstance(guidance_scale, List):
guidance_scale = [guidance_scale] * len(timesteps)
else:
guidance_scale = [
guidance_scale[guidance_mapping[i]] for i in range(len(timesteps))
]
if not isinstance(stg_scale, List):
stg_scale = [stg_scale] * len(timesteps)
else:
stg_scale = [stg_scale[guidance_mapping[i]] for i in range(len(timesteps))]
if not isinstance(rescaling_scale, List):
rescaling_scale = [rescaling_scale] * len(timesteps)
else:
rescaling_scale = [
rescaling_scale[guidance_mapping[i]] for i in range(len(timesteps))
]
# Normalize skip_block_list to always be None or a list of lists matching timesteps
if skip_block_list is not None:
# Convert single list to list of lists if needed
if len(skip_block_list) == 0 or not isinstance(skip_block_list[0], list):
skip_block_list = [skip_block_list] * len(timesteps)
else:
new_skip_block_list = []
for i, timestep in enumerate(timesteps):
new_skip_block_list.append(skip_block_list[guidance_mapping[i]])
skip_block_list = new_skip_block_list
if enhance_prompt:
self.prompt_enhancer_image_caption_model = (
self.prompt_enhancer_image_caption_model.to(self._execution_device)
)
self.prompt_enhancer_llm_model = self.prompt_enhancer_llm_model.to(
self._execution_device
)
prompt = generate_cinematic_prompt(
self.prompt_enhancer_image_caption_model,
self.prompt_enhancer_image_caption_processor,
self.prompt_enhancer_llm_model,
self.prompt_enhancer_llm_tokenizer,
prompt,
conditioning_items,
max_new_tokens=text_encoder_max_tokens,
)
# 3. Encode input prompt
if self.text_encoder is not None:
self.text_encoder = self.text_encoder.to(self._execution_device)
(
prompt_embeds,
prompt_attention_mask,
negative_prompt_embeds,
negative_prompt_attention_mask,
) = self.encode_prompt(
prompt,
True,
negative_prompt=negative_prompt,
num_images_per_prompt=num_images_per_prompt,
device=device,
prompt_embeds=prompt_embeds,
negative_prompt_embeds=negative_prompt_embeds,
prompt_attention_mask=prompt_attention_mask,
negative_prompt_attention_mask=negative_prompt_attention_mask,
text_encoder_max_tokens=text_encoder_max_tokens,
)
if offload_to_cpu and self.text_encoder is not None:
self.text_encoder = self.text_encoder.cpu()
self.transformer = self.transformer.to(self._execution_device)
prompt_embeds_batch = prompt_embeds
prompt_attention_mask_batch = prompt_attention_mask
negative_prompt_embeds = (
torch.zeros_like(prompt_embeds)
if negative_prompt_embeds is None
else negative_prompt_embeds
)
negative_prompt_attention_mask = (
torch.zeros_like(prompt_attention_mask)
if negative_prompt_attention_mask is None
else negative_prompt_attention_mask
)
prompt_embeds_batch = torch.cat(
[negative_prompt_embeds, prompt_embeds, prompt_embeds], dim=0
)
prompt_attention_mask_batch = torch.cat(
[
negative_prompt_attention_mask,
prompt_attention_mask,
prompt_attention_mask,
],
dim=0,
)
# 4. Prepare the initial latents using the provided media and conditioning items
# Prepare the initial latents tensor, shape = (b, c, f, h, w)
latents = self.prepare_latents(
latents=latents,
media_items=media_items,
timestep=timesteps[0],
latent_shape=latent_shape,
dtype=prompt_embeds.dtype,
device=device,
generator=generator,
vae_per_channel_normalize=vae_per_channel_normalize,
)
# Update the latents with the conditioning items and patchify them into (b, n, c)
latents, pixel_coords, conditioning_mask, num_cond_latents = (
self.prepare_conditioning(
conditioning_items=conditioning_items,
init_latents=latents,
num_frames=num_frames,
height=height,
width=width,
vae_per_channel_normalize=vae_per_channel_normalize,
generator=generator,
)
)
init_latents = latents.clone() # Used for image_cond_noise_update
# 6. Prepare extra step kwargs. TODO: Logic should ideally just be moved out of the pipeline
extra_step_kwargs = self.prepare_extra_step_kwargs(generator, eta)
# 7. Denoising loop
num_warmup_steps = max(
len(timesteps) - num_inference_steps * self.scheduler.order, 0
)
orig_conditioning_mask = conditioning_mask
# Befor compiling this code please be aware:
# This code might generate different input shapes if some timesteps have no STG or CFG.
# This means that the codes might need to be compiled mutliple times.
# To avoid that, use the same STG and CFG values for all timesteps.
with self.progress_bar(total=num_inference_steps) as progress_bar:
for i, t in enumerate(timesteps):
do_classifier_free_guidance = guidance_scale[i] > 1.0
do_spatio_temporal_guidance = stg_scale[i] > 0
do_rescaling = rescaling_scale[i] != 1.0
num_conds = 1
if do_classifier_free_guidance:
num_conds += 1
if do_spatio_temporal_guidance:
num_conds += 1
if do_classifier_free_guidance and do_spatio_temporal_guidance:
indices = slice(batch_size * 0, batch_size * 3)
elif do_classifier_free_guidance:
indices = slice(batch_size * 0, batch_size * 2)
elif do_spatio_temporal_guidance:
indices = slice(batch_size * 1, batch_size * 3)
else:
indices = slice(batch_size * 1, batch_size * 2)
# Prepare skip layer masks
skip_layer_mask: Optional[torch.Tensor] = None
if do_spatio_temporal_guidance:
if skip_block_list is not None:
skip_layer_mask = self.transformer.create_skip_layer_mask(
batch_size, num_conds, num_conds - 1, skip_block_list[i]
)
batch_pixel_coords = torch.cat([pixel_coords] * num_conds)
conditioning_mask = orig_conditioning_mask
if conditioning_mask is not None and is_video:
assert num_images_per_prompt == 1
conditioning_mask = torch.cat([conditioning_mask] * num_conds)
fractional_coords = batch_pixel_coords.to(torch.float32)
fractional_coords[:, 0] = fractional_coords[:, 0] * (1.0 / frame_rate)
if conditioning_mask is not None and image_cond_noise_scale > 0.0:
latents = self.add_noise_to_image_conditioning_latents(
t,
init_latents,
latents,
image_cond_noise_scale,
orig_conditioning_mask,
generator,
)
latent_model_input = (
torch.cat([latents] * num_conds) if num_conds > 1 else latents
)
latent_model_input = self.scheduler.scale_model_input(
latent_model_input, t
)
current_timestep = t
if not torch.is_tensor(current_timestep):
# TODO: this requires sync between CPU and GPU. So try to pass timesteps as tensors if you can
# This would be a good case for the `match` statement (Python 3.10+)
is_mps = latent_model_input.device.type == "mps"
if isinstance(current_timestep, float):
dtype = torch.float32 if is_mps else torch.float64
else:
dtype = torch.int32 if is_mps else torch.int64
current_timestep = torch.tensor(
[current_timestep],
dtype=dtype,
device=latent_model_input.device,
)
elif len(current_timestep.shape) == 0:
current_timestep = current_timestep[None].to(
latent_model_input.device
)
# broadcast to batch dimension in a way that's compatible with ONNX/Core ML
current_timestep = current_timestep.expand(
latent_model_input.shape[0]
).unsqueeze(-1)
if conditioning_mask is not None:
# Conditioning latents have an initial timestep and noising level of (1.0 - conditioning_mask)
# and will start to be denoised when the current timestep is lower than their conditioning timestep.
current_timestep = torch.min(
current_timestep, 1.0 - conditioning_mask
)
# Choose the appropriate context manager based on `mixed_precision`
if mixed_precision:
context_manager = torch.autocast(device.type, dtype=torch.bfloat16)
else:
context_manager = nullcontext() # Dummy context manager
# predict noise model_output
with context_manager:
noise_pred = self.transformer(
latent_model_input.to(self.transformer.dtype),
indices_grid=fractional_coords,
encoder_hidden_states=prompt_embeds_batch[indices].to(
self.transformer.dtype
),
encoder_attention_mask=prompt_attention_mask_batch[indices],
timestep=current_timestep,
skip_layer_mask=skip_layer_mask,
skip_layer_strategy=skip_layer_strategy,
return_dict=False,
)[0]
# perform guidance
if do_spatio_temporal_guidance:
noise_pred_text, noise_pred_text_perturb = noise_pred.chunk(
num_conds
)[-2:]
if do_classifier_free_guidance:
noise_pred_uncond, noise_pred_text = noise_pred.chunk(num_conds)[:2]
if cfg_star_rescale:
# Rescales the unconditional noise prediction using the projection of the conditional prediction onto it:
# α = (⟨ε_text, ε_uncond⟩ / ||ε_uncond||²), then ε_uncond ← α * ε_uncond
# where ε_text is the conditional noise prediction and ε_uncond is the unconditional one.
positive_flat = noise_pred_text.view(batch_size, -1)
negative_flat = noise_pred_uncond.view(batch_size, -1)
dot_product = torch.sum(
positive_flat * negative_flat, dim=1, keepdim=True
)
squared_norm = (
torch.sum(negative_flat**2, dim=1, keepdim=True) + 1e-8
)
alpha = dot_product / squared_norm
noise_pred_uncond = alpha * noise_pred_uncond
noise_pred = noise_pred_uncond + guidance_scale[i] * (
noise_pred_text - noise_pred_uncond
)
elif do_spatio_temporal_guidance:
noise_pred = noise_pred_text
if do_spatio_temporal_guidance:
noise_pred = noise_pred + stg_scale[i] * (
noise_pred_text - noise_pred_text_perturb
)
if do_rescaling and stg_scale[i] > 0.0:
noise_pred_text_std = noise_pred_text.view(batch_size, -1).std(
dim=1, keepdim=True
)
noise_pred_std = noise_pred.view(batch_size, -1).std(
dim=1, keepdim=True
)
factor = noise_pred_text_std / noise_pred_std
factor = rescaling_scale[i] * factor + (1 - rescaling_scale[i])
noise_pred = noise_pred * factor.view(batch_size, 1, 1)
current_timestep = current_timestep[:1]
# learned sigma
if (
self.transformer.config.out_channels // 2
== self.transformer.config.in_channels
):
noise_pred = noise_pred.chunk(2, dim=1)[0]
# compute previous image: x_t -> x_t-1
latents = self.denoising_step(
latents,
noise_pred,
current_timestep,
orig_conditioning_mask,
t,
extra_step_kwargs,
stochastic_sampling=stochastic_sampling,
)
# call the callback, if provided
if i == len(timesteps) - 1 or (
(i + 1) > num_warmup_steps and (i + 1) % self.scheduler.order == 0
):
progress_bar.update()
if callback_on_step_end is not None:
callback_on_step_end(self, i, t, {})
if offload_to_cpu:
self.transformer = self.transformer.cpu()
if self._execution_device == "cuda":
torch.cuda.empty_cache()
# Remove the added conditioning latents
latents = latents[:, num_cond_latents:]
latents = self.patchifier.unpatchify(
latents=latents,
output_height=latent_height,
output_width=latent_width,
out_channels=self.transformer.in_channels
// math.prod(self.patchifier.patch_size),
)
if output_type != "latent":
if self.vae.decoder.timestep_conditioning:
noise = torch.randn_like(latents)
if not isinstance(decode_timestep, list):
decode_timestep = [decode_timestep] * latents.shape[0]
if decode_noise_scale is None:
decode_noise_scale = decode_timestep
elif not isinstance(decode_noise_scale, list):
decode_noise_scale = [decode_noise_scale] * latents.shape[0]
decode_timestep = torch.tensor(decode_timestep).to(latents.device)
decode_noise_scale = torch.tensor(decode_noise_scale).to(
latents.device
)[:, None, None, None, None]
latents = (
latents * (1 - decode_noise_scale) + noise * decode_noise_scale
)
else:
decode_timestep = None
latents = self.tone_map_latents(latents, tone_map_compression_ratio)
image = vae_decode(
latents,
self.vae,
is_video,
vae_per_channel_normalize=kwargs["vae_per_channel_normalize"],
timestep=decode_timestep,
)
image = self.image_processor.postprocess(image, output_type=output_type)
else:
image = latents
# Offload all models
self.maybe_free_model_hooks()
if not return_dict:
return (image,)
return ImagePipelineOutput(images=image)
def denoising_step(
self,
latents: torch.Tensor,
noise_pred: torch.Tensor,
current_timestep: torch.Tensor,
conditioning_mask: torch.Tensor,
t: float,
extra_step_kwargs,
t_eps=1e-6,
stochastic_sampling=False,
):
"""
Perform the denoising step for the required tokens, based on the current timestep and
conditioning mask:
Conditioning latents have an initial timestep and noising level of (1.0 - conditioning_mask)
and will start to be denoised when the current timestep is equal or lower than their
conditioning timestep.
(hard-conditioning latents with conditioning_mask = 1.0 are never denoised)
"""
# Denoise the latents using the scheduler
denoised_latents = self.scheduler.step(
noise_pred,
t if current_timestep is None else current_timestep,
latents,
**extra_step_kwargs,
return_dict=False,
stochastic_sampling=stochastic_sampling,
)[0]
if conditioning_mask is None:
return denoised_latents
tokens_to_denoise_mask = (t - t_eps < (1.0 - conditioning_mask)).unsqueeze(-1)
return torch.where(tokens_to_denoise_mask, denoised_latents, latents)
def prepare_conditioning(
self,
conditioning_items: Optional[List[ConditioningItem]],
init_latents: torch.Tensor,
num_frames: int,
height: int,
width: int,
vae_per_channel_normalize: bool = False,
generator=None,
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor, int]:
"""
Prepare conditioning tokens based on the provided conditioning items.
This method encodes provided conditioning items (video frames or single frames) into latents
and integrates them with the initial latent tensor. It also calculates corresponding pixel
coordinates, a mask indicating the influence of conditioning latents, and the total number of
conditioning latents.
Args:
conditioning_items (Optional[List[ConditioningItem]]): A list of ConditioningItem objects.
init_latents (torch.Tensor): The initial latent tensor of shape (b, c, f_l, h_l, w_l), where
`f_l` is the number of latent frames, and `h_l` and `w_l` are latent spatial dimensions.
num_frames, height, width: The dimensions of the generated video.
vae_per_channel_normalize (bool, optional): Whether to normalize channels during VAE encoding.
Defaults to `False`.
generator: The random generator
Returns:
Tuple[torch.Tensor, torch.Tensor, torch.Tensor, int]:
- `init_latents` (torch.Tensor): The updated latent tensor including conditioning latents,
patchified into (b, n, c) shape.
- `init_pixel_coords` (torch.Tensor): The pixel coordinates corresponding to the updated
latent tensor.
- `conditioning_mask` (torch.Tensor): A mask indicating the conditioning-strength of each
latent token.
- `num_cond_latents` (int): The total number of latent tokens added from conditioning items.
Raises:
AssertionError: If input shapes, dimensions, or conditions for applying conditioning are invalid.
"""
assert isinstance(self.vae, CausalVideoAutoencoder)
if conditioning_items:
batch_size, _, num_latent_frames = init_latents.shape[:3]
init_conditioning_mask = torch.zeros(
init_latents[:, 0, :, :, :].shape,
dtype=torch.float32,
device=init_latents.device,
)
extra_conditioning_latents = []
extra_conditioning_pixel_coords = []
extra_conditioning_mask = []
extra_conditioning_num_latents = 0 # Number of extra conditioning latents added (should be removed before decoding)
# Process each conditioning item
for conditioning_item in conditioning_items:
conditioning_item = self._resize_conditioning_item(
conditioning_item, height, width
)
media_item = conditioning_item.media_item
media_frame_number = conditioning_item.media_frame_number
strength = conditioning_item.conditioning_strength
assert media_item.ndim == 5 # (b, c, f, h, w)
b, c, n_frames, h, w = media_item.shape
assert (
height == h and width == w
) or media_frame_number == 0, f"Dimensions do not match: {height}x{width} != {h}x{w} - allowed only when media_frame_number == 0"
assert n_frames % 8 == 1
assert (
media_frame_number >= 0
and media_frame_number + n_frames <= num_frames
)
# Encode the provided conditioning media item
media_item_latents = vae_encode(
media_item.to(dtype=self.vae.dtype, device=self.vae.device),
self.vae,
vae_per_channel_normalize=vae_per_channel_normalize,
).to(dtype=init_latents.dtype)
# Handle the different conditioning cases
if media_frame_number == 0:
# Get the target spatial position of the latent conditioning item
media_item_latents, l_x, l_y = self._get_latent_spatial_position(
media_item_latents,
conditioning_item,
height,
width,
strip_latent_border=True,
)
b, c_l, f_l, h_l, w_l = media_item_latents.shape
# First frame or sequence - just update the initial noise latents and the mask
init_latents[:, :, :f_l, l_y : l_y + h_l, l_x : l_x + w_l] = (
torch.lerp(
init_latents[:, :, :f_l, l_y : l_y + h_l, l_x : l_x + w_l],
media_item_latents,
strength,
)
)
init_conditioning_mask[
:, :f_l, l_y : l_y + h_l, l_x : l_x + w_l
] = strength
else:
# Non-first frame or sequence
if n_frames > 1:
# Handle non-first sequence.
# Encoded latents are either fully consumed, or the prefix is handled separately below.
(
init_latents,
init_conditioning_mask,
media_item_latents,
) = self._handle_non_first_conditioning_sequence(
init_latents,
init_conditioning_mask,
media_item_latents,
media_frame_number,
strength,
)
# Single frame or sequence-prefix latents
if media_item_latents is not None:
noise = randn_tensor(
media_item_latents.shape,
generator=generator,
device=media_item_latents.device,
dtype=media_item_latents.dtype,
)
media_item_latents = torch.lerp(
noise, media_item_latents, strength
)
# Patchify the extra conditioning latents and calculate their pixel coordinates
media_item_latents, latent_coords = self.patchifier.patchify(
latents=media_item_latents
)
pixel_coords = latent_to_pixel_coords(
latent_coords,
self.vae,
causal_fix=self.transformer.config.causal_temporal_positioning,
)
# Update the frame numbers to match the target frame number
pixel_coords[:, 0] += media_frame_number
extra_conditioning_num_latents += media_item_latents.shape[1]
conditioning_mask = torch.full(
media_item_latents.shape[:2],
strength,
dtype=torch.float32,
device=init_latents.device,
)
extra_conditioning_latents.append(media_item_latents)
extra_conditioning_pixel_coords.append(pixel_coords)
extra_conditioning_mask.append(conditioning_mask)
# Patchify the updated latents and calculate their pixel coordinates
init_latents, init_latent_coords = self.patchifier.patchify(
latents=init_latents
)
init_pixel_coords = latent_to_pixel_coords(
init_latent_coords,
self.vae,
causal_fix=self.transformer.config.causal_temporal_positioning,
)
if not conditioning_items:
return init_latents, init_pixel_coords, None, 0
init_conditioning_mask, _ = self.patchifier.patchify(
latents=init_conditioning_mask.unsqueeze(1)
)
init_conditioning_mask = init_conditioning_mask.squeeze(-1)
if extra_conditioning_latents:
# Stack the extra conditioning latents, pixel coordinates and mask
init_latents = torch.cat([*extra_conditioning_latents, init_latents], dim=1)
init_pixel_coords = torch.cat(
[*extra_conditioning_pixel_coords, init_pixel_coords], dim=2
)
init_conditioning_mask = torch.cat(
[*extra_conditioning_mask, init_conditioning_mask], dim=1
)
if self.transformer.use_tpu_flash_attention:
# When flash attention is used, keep the original number of tokens by removing
# tokens from the end.
init_latents = init_latents[:, :-extra_conditioning_num_latents]
init_pixel_coords = init_pixel_coords[
:, :, :-extra_conditioning_num_latents
]
init_conditioning_mask = init_conditioning_mask[
:, :-extra_conditioning_num_latents
]
return (
init_latents,
init_pixel_coords,
init_conditioning_mask,
extra_conditioning_num_latents,
)
@staticmethod
def _resize_conditioning_item(
conditioning_item: ConditioningItem,
height: int,
width: int,
):
if conditioning_item.media_x or conditioning_item.media_y:
raise ValueError(
"Provide media_item in the target size for spatial conditioning."
)
new_conditioning_item = copy.copy(conditioning_item)
new_conditioning_item.media_item = LTXVideoPipeline.resize_tensor(
conditioning_item.media_item, height, width
)
return new_conditioning_item
def _get_latent_spatial_position(
self,
latents: torch.Tensor,
conditioning_item: ConditioningItem,
height: int,
width: int,
strip_latent_border,
):
"""
Get the spatial position of the conditioning item in the latent space.
If requested, strip the conditioning latent borders that do not align with target borders.
(border latents look different then other latents and might confuse the model)
"""
scale = self.vae_scale_factor
h, w = conditioning_item.media_item.shape[-2:]
assert (
h <= height and w <= width
), f"Conditioning item size {h}x{w} is larger than target size {height}x{width}"
assert h % scale == 0 and w % scale == 0
# Compute the start and end spatial positions of the media item
x_start, y_start = conditioning_item.media_x, conditioning_item.media_y
x_start = (width - w) // 2 if x_start is None else x_start
y_start = (height - h) // 2 if y_start is None else y_start
x_end, y_end = x_start + w, y_start + h
assert (
x_end <= width and y_end <= height
), f"Conditioning item {x_start}:{x_end}x{y_start}:{y_end} is out of bounds for target size {width}x{height}"
if strip_latent_border:
# Strip one latent from left/right and/or top/bottom, update x, y accordingly
if x_start > 0:
x_start += scale
latents = latents[:, :, :, :, 1:]
if y_start > 0:
y_start += scale
latents = latents[:, :, :, 1:, :]
if x_end < width:
latents = latents[:, :, :, :, :-1]
if y_end < height:
latents = latents[:, :, :, :-1, :]
return latents, x_start // scale, y_start // scale
@staticmethod
def _handle_non_first_conditioning_sequence(
init_latents: torch.Tensor,
init_conditioning_mask: torch.Tensor,
latents: torch.Tensor,
media_frame_number: int,
strength: float,
num_prefix_latent_frames: int = 2,
prefix_latents_mode: str = "concat",
prefix_soft_conditioning_strength: float = 0.15,
):
"""
Special handling for a conditioning sequence that does not start on the first frame.
The special handling is required to allow a short encoded video to be used as middle
(or last) sequence in a longer video.
Args:
init_latents (torch.Tensor): The initial noise latents to be updated.
init_conditioning_mask (torch.Tensor): The initial conditioning mask to be updated.
latents (torch.Tensor): The encoded conditioning item.
media_frame_number (int): The target frame number of the first frame in the conditioning sequence.
strength (float): The conditioning strength for the conditioning latents.
num_prefix_latent_frames (int, optional): The length of the sequence prefix, to be handled
separately. Defaults to 2.
prefix_latents_mode (str, optional): Special treatment for prefix (boundary) latents.
- "drop": Drop the prefix latents.
- "soft": Use the prefix latents, but with soft-conditioning
- "concat": Add the prefix latents as extra tokens (like single frames)
prefix_soft_conditioning_strength (float, optional): The strength of the soft-conditioning for
the prefix latents, relevant if `prefix_latents_mode` is "soft". Defaults to 0.1.
"""
f_l = latents.shape[2]
f_l_p = num_prefix_latent_frames
assert f_l >= f_l_p
assert media_frame_number % 8 == 0
if f_l > f_l_p:
# Insert the conditioning latents **excluding the prefix** into the sequence
f_l_start = media_frame_number // 8 + f_l_p
f_l_end = f_l_start + f_l - f_l_p
init_latents[:, :, f_l_start:f_l_end] = torch.lerp(
init_latents[:, :, f_l_start:f_l_end],
latents[:, :, f_l_p:],
strength,
)
# Mark these latent frames as conditioning latents
init_conditioning_mask[:, f_l_start:f_l_end] = strength
# Handle the prefix-latents
if prefix_latents_mode == "soft":
if f_l_p > 1:
# Drop the first (single-frame) latent and soft-condition the remaining prefix
f_l_start = media_frame_number // 8 + 1
f_l_end = f_l_start + f_l_p - 1
strength = min(prefix_soft_conditioning_strength, strength)
init_latents[:, :, f_l_start:f_l_end] = torch.lerp(
init_latents[:, :, f_l_start:f_l_end],
latents[:, :, 1:f_l_p],
strength,
)
# Mark these latent frames as conditioning latents
init_conditioning_mask[:, f_l_start:f_l_end] = strength
latents = None # No more latents to handle
elif prefix_latents_mode == "drop":
# Drop the prefix latents
latents = None
elif prefix_latents_mode == "concat":
# Pass-on the prefix latents to be handled as extra conditioning frames
latents = latents[:, :, :f_l_p]
else:
raise ValueError(f"Invalid prefix_latents_mode: {prefix_latents_mode}")
return (
init_latents,
init_conditioning_mask,
latents,
)
def trim_conditioning_sequence(
self, start_frame: int, sequence_num_frames: int, target_num_frames: int
):
"""
Trim a conditioning sequence to the allowed number of frames.
Args:
start_frame (int): The target frame number of the first frame in the sequence.
sequence_num_frames (int): The number of frames in the sequence.
target_num_frames (int): The target number of frames in the generated video.
Returns:
int: updated sequence length
"""
scale_factor = self.video_scale_factor
num_frames = min(sequence_num_frames, target_num_frames - start_frame)
# Trim down to a multiple of temporal_scale_factor frames plus 1
num_frames = (num_frames - 1) // scale_factor * scale_factor + 1
return num_frames
@staticmethod
def tone_map_latents(
latents: torch.Tensor,
compression: float,
) -> torch.Tensor:
"""
Applies a non-linear tone-mapping function to latent values to reduce their dynamic range
in a perceptually smooth way using a sigmoid-based compression.
This is useful for regularizing high-variance latents or for conditioning outputs
during generation, especially when controlling dynamic behavior with a `compression` factor.
Parameters:
----------
latents : torch.Tensor
Input latent tensor with arbitrary shape. Expected to be roughly in [-1, 1] or [0, 1] range.
compression : float
Compression strength in the range [0, 1].
- 0.0: No tone-mapping (identity transform)
- 1.0: Full compression effect
Returns:
-------
torch.Tensor
The tone-mapped latent tensor of the same shape as input.
"""
if not (0 <= compression <= 1):
raise ValueError("Compression must be in the range [0, 1]")
# Remap [0-1] to [0-0.75] and apply sigmoid compression in one shot
scale_factor = compression * 0.75
abs_latents = torch.abs(latents)
# Sigmoid compression: sigmoid shifts large values toward 0.2, small values stay ~1.0
# When scale_factor=0, sigmoid term vanishes, when scale_factor=0.75, full effect
sigmoid_term = torch.sigmoid(4.0 * scale_factor * (abs_latents - 1.0))
scales = 1.0 - 0.8 * scale_factor * sigmoid_term
filtered = latents * scales
return filtered
def adain_filter_latent(
latents: torch.Tensor, reference_latents: torch.Tensor, factor=1.0
):
"""
Applies Adaptive Instance Normalization (AdaIN) to a latent tensor based on
statistics from a reference latent tensor.
Args:
latent (torch.Tensor): Input latents to normalize
reference_latent (torch.Tensor): The reference latents providing style statistics.
factor (float): Blending factor between original and transformed latent.
Range: -10.0 to 10.0, Default: 1.0
Returns:
torch.Tensor: The transformed latent tensor
"""
result = latents.clone()
for i in range(latents.size(0)):
for c in range(latents.size(1)):
r_sd, r_mean = torch.std_mean(
reference_latents[i, c], dim=None
) # index by original dim order
i_sd, i_mean = torch.std_mean(result[i, c], dim=None)
result[i, c] = ((result[i, c] - i_mean) / i_sd) * r_sd + r_mean
result = torch.lerp(latents, result, factor)
return result
class LTXMultiScalePipeline:
def _upsample_latents(
self, latest_upsampler: LatentUpsampler, latents: torch.Tensor
):
assert latents.device == latest_upsampler.device
latents = un_normalize_latents(
latents, self.vae, vae_per_channel_normalize=True
)
upsampled_latents = latest_upsampler(latents)
upsampled_latents = normalize_latents(
upsampled_latents, self.vae, vae_per_channel_normalize=True
)
return upsampled_latents
def __init__(
self, video_pipeline: LTXVideoPipeline, latent_upsampler: LatentUpsampler
):
self.video_pipeline = video_pipeline
self.vae = video_pipeline.vae
self.latent_upsampler = latent_upsampler
def __call__(
self,
downscale_factor: float,
first_pass: dict,
second_pass: dict,
*args: Any,
**kwargs: Any,
) -> Any:
original_kwargs = kwargs.copy()
original_output_type = kwargs["output_type"]
original_width = kwargs["width"]
original_height = kwargs["height"]
x_width = int(kwargs["width"] * downscale_factor)
downscaled_width = x_width - (x_width % self.video_pipeline.vae_scale_factor)
x_height = int(kwargs["height"] * downscale_factor)
downscaled_height = x_height - (x_height % self.video_pipeline.vae_scale_factor)
kwargs["output_type"] = "latent"
kwargs["width"] = downscaled_width
kwargs["height"] = downscaled_height
kwargs.update(**first_pass)
result = self.video_pipeline(*args, **kwargs)
latents = result.images
upsampled_latents = self._upsample_latents(self.latent_upsampler, latents)
upsampled_latents = adain_filter_latent(
latents=upsampled_latents, reference_latents=latents
)
kwargs = original_kwargs
kwargs["latents"] = upsampled_latents
kwargs["output_type"] = original_output_type
kwargs["width"] = downscaled_width * 2
kwargs["height"] = downscaled_height * 2
kwargs.update(**second_pass)
result = self.video_pipeline(*args, **kwargs)
if original_output_type != "latent":
num_frames = result.images.shape[2]
videos = rearrange(result.images, "b c f h w -> (b f) c h w")
videos = F.interpolate(
videos,
size=(original_height, original_width),
mode="bilinear",
align_corners=False,
)
videos = rearrange(videos, "(b f) c h w -> b c f h w", f=num_frames)
result.images = videos
return result
================================================
FILE: ltx_video/schedulers/__init__.py
================================================
================================================
FILE: ltx_video/schedulers/rf.py
================================================
import math
from abc import ABC, abstractmethod
from dataclasses import dataclass
from typing import Callable, Optional, Tuple, Union
import json
import os
from pathlib import Path
import torch
from diffusers.configuration_utils import ConfigMixin, register_to_config
from diffusers.schedulers.scheduling_utils import SchedulerMixin
from diffusers.utils import BaseOutput
from torch import Tensor
from safetensors import safe_open
from ltx_video.utils.torch_utils import append_dims
from ltx_video.utils.diffusers_config_mapping import (
diffusers_and_ours_config_mapping,
make_hashable_key,
)
def linear_quadratic_schedule(num_steps, threshold_noise=0.025, linear_steps=None):
if num_steps == 1:
return torch.tensor([1.0])
if linear_steps is None:
linear_steps = num_steps // 2
linear_sigma_schedule = [
i * threshold_noise / linear_steps for i in range(linear_steps)
]
threshold_noise_step_diff = linear_steps - threshold_noise * num_steps
quadratic_steps = num_steps - linear_steps
quadratic_coef = threshold_noise_step_diff / (linear_steps * quadratic_steps**2)
linear_coef = threshold_noise / linear_steps - 2 * threshold_noise_step_diff / (
quadratic_steps**2
)
const = quadratic_coef * (linear_steps**2)
quadratic_sigma_schedule = [
quadratic_coef * (i**2) + linear_coef * i + const
for i in range(linear_steps, num_steps)
]
sigma_schedule = linear_sigma_schedule + quadratic_sigma_schedule + [1.0]
sigma_schedule = [1.0 - x for x in sigma_schedule]
return torch.tensor(sigma_schedule[:-1])
def simple_diffusion_resolution_dependent_timestep_shift(
samples_shape: torch.Size,
timesteps: Tensor,
n: int = 32 * 32,
) -> Tensor:
if len(samples_shape) == 3:
_, m, _ = samples_shape
elif len(samples_shape) in [4, 5]:
m = math.prod(samples_shape[2:])
else:
raise ValueError(
"Samples must have shape (b, t, c), (b, c, h, w) or (b, c, f, h, w)"
)
snr = (timesteps / (1 - timesteps)) ** 2
shift_snr = torch.log(snr) + 2 * math.log(m / n)
shifted_timesteps = torch.sigmoid(0.5 * shift_snr)
return shifted_timesteps
def time_shift(mu: float, sigma: float, t: Tensor):
return math.exp(mu) / (math.exp(mu) + (1 / t - 1) ** sigma)
def get_normal_shift(
n_tokens: int,
min_tokens: int = 1024,
max_tokens: int = 4096,
min_shift: float = 0.95,
max_shift: float = 2.05,
) -> Callable[[float], float]:
m = (max_shift - min_shift) / (max_tokens - min_tokens)
b = min_shift - m * min_tokens
return m * n_tokens + b
def strech_shifts_to_terminal(shifts: Tensor, terminal=0.1):
"""
Stretch a function (given as sampled shifts) so that its final value matches the given terminal value
using the provided formula.
Parameters:
- shifts (Tensor): The samples of the function to be stretched (PyTorch Tensor).
- terminal (float): The desired terminal value (value at the last sample).
Returns:
- Tensor: The stretched shifts such that the final value equals `terminal`.
"""
if shifts.numel() == 0:
raise ValueError("The 'shifts' tensor must not be empty.")
# Ensure terminal value is valid
if terminal <= 0 or terminal >= 1:
raise ValueError("The terminal value must be between 0 and 1 (exclusive).")
# Transform the shifts using the given formula
one_minus_z = 1 - shifts
scale_factor = one_minus_z[-1] / (1 - terminal)
stretched_shifts = 1 - (one_minus_z / scale_factor)
return stretched_shifts
def sd3_resolution_dependent_timestep_shift(
samples_shape: torch.Size,
timesteps: Tensor,
target_shift_terminal: Optional[float] = None,
) -> Tensor:
"""
Shifts the timestep schedule as a function of the generated resolution.
In the SD3 paper, the authors empirically how to shift the timesteps based on the resolution of the target images.
For more details: https://arxiv.org/pdf/2403.03206
In Flux they later propose a more dynamic resolution dependent timestep shift, see:
https://github.com/black-forest-labs/flux/blob/87f6fff727a377ea1c378af692afb41ae84cbe04/src/flux/sampling.py#L66
Args:
samples_shape (torch.Size): The samples batch shape (batch_size, channels, height, width) or
(batch_size, channels, frame, height, width).
timesteps (Tensor): A batch of timesteps with shape (batch_size,).
target_shift_terminal (float): The target terminal value for the shifted timesteps.
Returns:
Tensor: The shifted timesteps.
"""
if len(samples_shape) == 3:
_, m, _ = samples_shape
elif len(samples_shape) in [4, 5]:
m = math.prod(samples_shape[2:])
else:
raise ValueError(
"Samples must have shape (b, t, c), (b, c, h, w) or (b, c, f, h, w)"
)
shift = get_normal_shift(m)
time_shifts = time_shift(shift, 1, timesteps)
if target_shift_terminal is not None: # Stretch the shifts to the target terminal
time_shifts = strech_shifts_to_terminal(time_shifts, target_shift_terminal)
return time_shifts
class TimestepShifter(ABC):
@abstractmethod
def shift_timesteps(self, samples_shape: torch.Size, timesteps: Tensor) -> Tensor:
pass
@dataclass
class RectifiedFlowSchedulerOutput(BaseOutput):
"""
Output class for the scheduler's step function output.
Args:
prev_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
Computed sample (x_{t-1}) of previous timestep. `prev_sample` should be used as next model input in the
denoising loop.
pred_original_sample (`torch.FloatTensor` of shape `(batch_size, num_channels, height, width)` for images):
The predicted denoised sample (x_{0}) based on the model output from the current timestep.
`pred_original_sample` can be used to preview progress or for guidance.
"""
prev_sample: torch.FloatTensor
pred_original_sample: Optional[torch.FloatTensor] = None
class RectifiedFlowScheduler(SchedulerMixin, ConfigMixin, TimestepShifter):
order = 1
@register_to_config
def __init__(
self,
num_train_timesteps=1000,
shifting: Optional[str] = None,
base_resolution: int = 32**2,
target_shift_terminal: Optional[float] = None,
sampler: Optional[str] = "Uniform",
shift: Optional[float] = None,
):
super().__init__()
self.init_noise_sigma = 1.0
self.num_inference_steps = None
self.sampler = sampler
self.shifting = shifting
self.base_resolution = base_resolution
self.target_shift_terminal = target_shift_terminal
self.timesteps = self.sigmas = self.get_initial_timesteps(
num_train_timesteps, shift=shift
)
self.shift = shift
def get_initial_timesteps(
self, num_timesteps: int, shift: Optional[float] = None
) -> Tensor:
if self.sampler == "Uniform":
return torch.linspace(1, 1 / num_timesteps, num_timesteps)
elif self.sampler == "LinearQuadratic":
return linear_quadratic_schedule(num_timesteps)
elif self.sampler == "Constant":
assert (
shift is not None
), "Shift must be provided for constant time shift sampler."
return time_shift(
shift, 1, torch.linspace(1, 1 / num_timesteps, num_timesteps)
)
def shift_timesteps(self, samples_shape: torch.Size, timesteps: Tensor) -> Tensor:
if self.shifting == "SD3":
return sd3_resolution_dependent_timestep_shift(
samples_shape, timesteps, self.target_shift_terminal
)
elif self.shifting == "SimpleDiffusion":
return simple_diffusion_resolution_dependent_timestep_shift(
samples_shape, timesteps, self.base_resolution
)
return timesteps
def set_timesteps(
self,
num_inference_steps: Optional[int] = None,
samples_shape: Optional[torch.Size] = None,
timesteps: Optional[Tensor] = None,
device: Union[str, torch.device] = None,
):
"""
Sets the discrete timesteps used for the diffusion chain. Supporting function to be run before inference.
If `timesteps` are provided, they will be used instead of the scheduled timesteps.
Args:
num_inference_steps (`int` *optional*): The number of diffusion steps used when generating samples.
samples_shape (`torch.Size` *optional*): The samples batch shape, used for shifting.
timesteps ('torch.Tensor' *optional*): Specific timesteps to use instead of scheduled timesteps.
device (`Union[str, torch.device]`, *optional*): The device to which the timesteps tensor will be moved.
"""
if timesteps is not None and num_inference_steps is not None:
raise ValueError(
"You cannot provide both `timesteps` and `num_inference_steps`."
)
if timesteps is None:
num_inference_steps = min(
self.config.num_train_timesteps, num_inference_steps
)
timesteps = self.get_initial_timesteps(
num_inference_steps, shift=self.shift
).to(device)
timesteps = self.shift_timesteps(samples_shape, timesteps)
else:
timesteps = torch.Tensor(timesteps).to(device)
num_inference_steps = len(timesteps)
self.timesteps = timesteps
self.num_inference_steps = num_inference_steps
self.sigmas = self.timesteps
@staticmethod
def from_pretrained(pretrained_model_path: Union[str, os.PathLike]):
pretrained_model_path = Path(pretrained_model_path)
if pretrained_model_path.is_file():
comfy_single_file_state_dict = {}
with safe_open(pretrained_model_path, framework="pt", device="cpu") as f:
metadata = f.metadata()
for k in f.keys():
comfy_single_file_state_dict[k] = f.get_tensor(k)
configs = json.loads(metadata["config"])
config = configs["scheduler"]
del comfy_single_file_state_dict
elif pretrained_model_path.is_dir():
diffusers_noise_scheduler_config_path = (
pretrained_model_path / "scheduler" / "scheduler_config.json"
)
with open(diffusers_noise_scheduler_config_path, "r") as f:
scheduler_config = json.load(f)
hashable_config = make_hashable_key(scheduler_config)
if hashable_config in diffusers_and_ours_config_mapping:
config = diffusers_and_ours_config_mapping[hashable_config]
return RectifiedFlowScheduler.from_config(config)
def scale_model_input(
self, sample: torch.FloatTensor, timestep: Optional[int] = None
) -> torch.FloatTensor:
# pylint: disable=unused-argument
"""
Ensures interchangeability with schedulers that need to scale the denoising model input depending on the
current timestep.
Args:
sample (`torch.FloatTensor`): input sample
timestep (`int`, optional): current timestep
Returns:
`torch.FloatTensor`: scaled input sample
"""
return sample
def step(
self,
model_output: torch.FloatTensor,
timestep: torch.FloatTensor,
sample: torch.FloatTensor,
return_dict: bool = True,
stochastic_sampling: Optional[bool] = False,
**kwargs,
) -> Union[RectifiedFlowSchedulerOutput, Tuple]:
"""
Predict the sample from the previous timestep by reversing the SDE. This function propagates the diffusion
process from the learned model outputs (most often the predicted noise).
z_{t_1} = z_t - Delta_t * v
The method finds the next timestep that is lower than the input timestep(s) and denoises the latents
to that level. The input timestep(s) are not required to be one of the predefined timesteps.
Args:
model_output (`torch.FloatTensor`):
The direct output from learned diffusion model - the velocity,
timestep (`float`):
The current discrete timestep in the diffusion chain (global or per-token).
sample (`torch.FloatTensor`):
A current latent tokens to be de-noised.
return_dict (`bool`, *optional*, defaults to `True`):
Whether or not to return a [`~schedulers.scheduling_ddim.DDIMSchedulerOutput`] or `tuple`.
stochastic_sampling (`bool`, *optional*, defaults to `False`):
Whether to use stochastic sampling for the sampling process.
Returns:
[`~schedulers.scheduling_utils.RectifiedFlowSchedulerOutput`] or `tuple`:
If return_dict is `True`, [`~schedulers.rf_scheduler.RectifiedFlowSchedulerOutput`] is returned,
otherwise a tuple is returned where the first element is the sample tensor.
"""
if self.num_inference_steps is None:
raise ValueError(
"Number of inference steps is 'None', you need to run 'set_timesteps' after creating the scheduler"
)
t_eps = 1e-6 # Small epsilon to avoid numerical issues in timestep values
timesteps_padded = torch.cat(
[self.timesteps, torch.zeros(1, device=self.timesteps.device)]
)
# Find the next lower timestep(s) and compute the dt from the current timestep(s)
if timestep.ndim == 0:
# Global timestep case
lower_mask = timesteps_padded < timestep - t_eps
lower_timestep = timesteps_padded[lower_mask][0] # Closest lower timestep
dt = timestep - lower_timestep
else:
# Per-token case
assert timestep.ndim == 2
lower_mask = timesteps_padded[:, None, None] < timestep[None] - t_eps
lower_timestep = lower_mask * timesteps_padded[:, None, None]
lower_timestep, _ = lower_timestep.max(dim=0)
dt = (timestep - lower_timestep)[..., None]
# Compute previous sample
if stochastic_sampling:
x0 = sample - timestep[..., None] * model_output
next_timestep = timestep[..., None] - dt
prev_sample = self.add_noise(x0, torch.randn_like(sample), next_timestep)
else:
prev_sample = sample - dt * model_output
if not return_dict:
return (prev_sample,)
return RectifiedFlowSchedulerOutput(prev_sample=prev_sample)
def add_noise(
self,
original_samples: torch.FloatTensor,
noise: torch.FloatTensor,
timesteps: torch.FloatTensor,
) -> torch.FloatTensor:
sigmas = timesteps
sigmas = append_dims(sigmas, original_samples.ndim)
alphas = 1 - sigmas
noisy_samples = alphas * original_samples + sigmas * noise
return noisy_samples
================================================
FILE: ltx_video/utils/__init__.py
================================================
================================================
FILE: ltx_video/utils/diffusers_config_mapping.py
================================================
def make_hashable_key(dict_key):
def convert_value(value):
if isinstance(value, list):
return tuple(value)
elif isinstance(value, dict):
return tuple(sorted((k, convert_value(v)) for k, v in value.items()))
else:
return value
return tuple(sorted((k, convert_value(v)) for k, v in dict_key.items()))
DIFFUSERS_SCHEDULER_CONFIG = {
"_class_name": "FlowMatchEulerDiscreteScheduler",
"_diffusers_version": "0.32.0.dev0",
"base_image_seq_len": 1024,
"base_shift": 0.95,
"invert_sigmas": False,
"max_image_seq_len": 4096,
"max_shift": 2.05,
"num_train_timesteps": 1000,
"shift": 1.0,
"shift_terminal": 0.1,
"use_beta_sigmas": False,
"use_dynamic_shifting": True,
"use_exponential_sigmas": False,
"use_karras_sigmas": False,
}
DIFFUSERS_TRANSFORMER_CONFIG = {
"_class_name": "LTXVideoTransformer3DModel",
"_diffusers_version": "0.32.0.dev0",
"activation_fn": "gelu-approximate",
"attention_bias": True,
"attention_head_dim": 64,
"attention_out_bias": True,
"caption_channels": 4096,
"cross_attention_dim": 2048,
"in_channels": 128,
"norm_elementwise_affine": False,
"norm_eps": 1e-06,
"num_attention_heads": 32,
"num_layers": 28,
"out_channels": 128,
"patch_size": 1,
"patch_size_t": 1,
"qk_norm": "rms_norm_across_heads",
}
DIFFUSERS_VAE_CONFIG = {
"_class_name": "AutoencoderKLLTXVideo",
"_diffusers_version": "0.32.0.dev0",
"block_out_channels": [128, 256, 512, 512],
"decoder_causal": False,
"encoder_causal": True,
"in_channels": 3,
"latent_channels": 128,
"layers_per_block": [4, 3, 3, 3, 4],
"out_channels": 3,
"patch_size": 4,
"patch_size_t": 1,
"resnet_norm_eps": 1e-06,
"scaling_factor": 1.0,
"spatio_temporal_scaling": [True, True, True, False],
}
OURS_SCHEDULER_CONFIG = {
"_class_name": "RectifiedFlowScheduler",
"_diffusers_version": "0.25.1",
"num_train_timesteps": 1000,
"shifting": "SD3",
"base_resolution": None,
"target_shift_terminal": 0.1,
}
OURS_TRANSFORMER_CONFIG = {
"_class_name": "Transformer3DModel",
"_diffusers_version": "0.25.1",
"_name_or_path": "PixArt-alpha/PixArt-XL-2-256x256",
"activation_fn": "gelu-approximate",
"attention_bias": True,
"attention_head_dim": 64,
"attention_type": "default",
"caption_channels": 4096,
"cross_attention_dim": 2048,
"double_self_attention": False,
"dropout": 0.0,
"in_channels": 128,
"norm_elementwise_affine": False,
"norm_eps": 1e-06,
"norm_num_groups": 32,
"num_attention_heads": 32,
"num_embeds_ada_norm": 1000,
"num_layers": 28,
"num_vector_embeds": None,
"only_cross_attention": False,
"out_channels": 128,
"project_to_2d_pos": True,
"upcast_attention": False,
"use_linear_projection": False,
"qk_norm": "rms_norm",
"standardization_norm": "rms_norm",
"positional_embedding_type": "rope",
"positional_embedding_theta": 10000.0,
"positional_embedding_max_pos": [20, 2048, 2048],
"timestep_scale_multiplier": 1000,
}
OURS_VAE_CONFIG = {
"_class_name": "CausalVideoAutoencoder",
"dims": 3,
"in_channels": 3,
"out_channels": 3,
"latent_channels": 128,
"blocks": [
["res_x", 4],
["compress_all", 1],
["res_x_y", 1],
["res_x", 3],
["compress_all", 1],
["res_x_y", 1],
["res_x", 3],
["compress_all", 1],
["res_x", 3],
["res_x", 4],
],
"scaling_factor": 1.0,
"norm_layer": "pixel_norm",
"patch_size": 4,
"latent_log_var": "uniform",
"use_quant_conv": False,
"causal_decoder": False,
}
diffusers_and_ours_config_mapping = {
make_hashable_key(DIFFUSERS_SCHEDULER_CONFIG): OURS_SCHEDULER_CONFIG,
make_hashable_key(DIFFUSERS_TRANSFORMER_CONFIG): OURS_TRANSFORMER_CONFIG,
make_hashable_key(DIFFUSERS_VAE_CONFIG): OURS_VAE_CONFIG,
}
TRANSFORMER_KEYS_RENAME_DICT = {
"proj_in": "patchify_proj",
"time_embed": "adaln_single",
"norm_q": "q_norm",
"norm_k": "k_norm",
}
VAE_KEYS_RENAME_DICT = {
"decoder.up_blocks.3.conv_in": "decoder.up_blocks.7",
"decoder.up_blocks.3.upsamplers.0": "decoder.up_blocks.8",
"decoder.up_blocks.3": "decoder.up_blocks.9",
"decoder.up_blocks.2.upsamplers.0": "decoder.up_blocks.5",
"decoder.up_blocks.2.conv_in": "decoder.up_blocks.4",
"decoder.up_blocks.2": "decoder.up_blocks.6",
"decoder.up_blocks.1.upsamplers.0": "decoder.up_blocks.2",
"decoder.up_blocks.1": "decoder.up_blocks.3",
"decoder.up_blocks.0": "decoder.up_blocks.1",
"decoder.mid_block": "decoder.up_blocks.0",
"encoder.down_blocks.3": "encoder.down_blocks.8",
"encoder.down_blocks.2.downsamplers.0": "encoder.down_blocks.7",
"encoder.down_blocks.2": "encoder.down_blocks.6",
"encoder.down_blocks.1.downsamplers.0": "encoder.down_blocks.4",
"encoder.down_blocks.1.conv_out": "encoder.down_blocks.5",
"encoder.down_blocks.1": "encoder.down_blocks.3",
"encoder.down_blocks.0.conv_out": "encoder.down_blocks.2",
"encoder.down_blocks.0.downsamplers.0": "encoder.down_blocks.1",
"encoder.down_blocks.0": "encoder.down_blocks.0",
"encoder.mid_block": "encoder.down_blocks.9",
"conv_shortcut.conv": "conv_shortcut",
"resnets": "res_blocks",
"norm3": "norm3.norm",
"latents_mean": "per_channel_statistics.mean-of-means",
"latents_std": "per_channel_statistics.std-of-means",
}
================================================
FILE: ltx_video/utils/prompt_enhance_utils.py
================================================
import logging
from typing import Union, List, Optional
import torch
from PIL import Image
logger = logging.getLogger(__name__) # pylint: disable=invalid-name
T2V_CINEMATIC_PROMPT = """You are an expert cinematic director with many award winning movies, When writing prompts based on the user input, focus on detailed, chronological descriptions of actions and scenes.
Include specific movements, appearances, camera angles, and environmental details - all in a single flowing paragraph.
Start directly with the action, and keep descriptions literal and precise.
Think like a cinematographer describing a shot list.
Do not change the user input intent, just enhance it.
Keep within 150 words.
For best results, build your prompts using this structure:
Start with main action in a single sentence
Add specific details about movements and gestures
Describe character/object appearances precisely
Include background and environment details
Specify camera angles and movements
Describe lighting and colors
Note any changes or sudden events
Do not exceed the 150 word limit!
Output the enhanced prompt only.
"""
I2V_CINEMATIC_PROMPT = """You are an expert cinematic director with many award winning movies, When writing prompts based on the user input, focus on detailed, chronological descriptions of actions and scenes.
Include specific movements, appearances, camera angles, and environmental details - all in a single flowing paragraph.
Start directly with the action, and keep descriptions literal and precise.
Think like a cinematographer describing a shot list.
Keep within 150 words.
For best results, build your prompts using this structure:
Describe the image first and then add the user input. Image description should be in first priority! Align to the image caption if it contradicts the user text input.
Start with main action in a single sentence
Add specific details about movements and gestures
Describe character/object appearances precisely
Include background and environment details
Specify camera angles and movements
Describe lighting and colors
Note any changes or sudden events
Align to the image caption if it contradicts the user text input.
Do not exceed the 150 word limit!
Output the enhanced prompt only.
"""
def tensor_to_pil(tensor):
# Ensure tensor is in range [-1, 1]
assert tensor.min() >= -1 and tensor.max() <= 1
# Convert from [-1, 1] to [0, 1]
tensor = (tensor + 1) / 2
# Rearrange from [C, H, W] to [H, W, C]
tensor = tensor.permute(1, 2, 0)
# Convert to numpy array and then to uint8 range [0, 255]
numpy_image = (tensor.cpu().numpy() * 255).astype("uint8")
# Convert to PIL Image
return Image.fromarray(numpy_image)
def generate_cinematic_prompt(
image_caption_model,
image_caption_processor,
prompt_enhancer_model,
prompt_enhancer_tokenizer,
prompt: Union[str, List[str]],
conditioning_items: Optional[List] = None,
max_new_tokens: int = 256,
) -> List[str]:
prompts = [prompt] if isinstance(prompt, str) else prompt
if conditioning_items is None:
prompts = _generate_t2v_prompt(
prompt_enhancer_model,
prompt_enhancer_tokenizer,
prompts,
max_new_tokens,
T2V_CINEMATIC_PROMPT,
)
else:
if len(conditioning_items) > 1 or conditioning_items[0].media_frame_number != 0:
logger.warning(
"prompt enhancement does only support unconditional or first frame of conditioning items, returning original prompts"
)
return prompts
first_frame_conditioning_item = conditioning_items[0]
first_frames = _get_first_frames_from_conditioning_item(
first_frame_conditioning_item
)
assert len(first_frames) == len(
prompts
), "Number of conditioning frames must match number of prompts"
prompts = _generate_i2v_prompt(
image_caption_model,
image_caption_processor,
prompt_enhancer_model,
prompt_enhancer_tokenizer,
prompts,
first_frames,
max_new_tokens,
I2V_CINEMATIC_PROMPT,
)
return prompts
def _get_first_frames_from_conditioning_item(conditioning_item) -> List[Image.Image]:
frames_tensor = conditioning_item.media_item
return [
tensor_to_pil(frames_tensor[i, :, 0, :, :])
for i in range(frames_tensor.shape[0])
]
def _generate_t2v_prompt(
prompt_enhancer_model,
prompt_enhancer_tokenizer,
prompts: List[str],
max_new_tokens: int,
system_prompt: str,
) -> List[str]:
messages = [
[
{"role": "system", "content": system_prompt},
{"role": "user", "content": f"user_prompt: {p}"},
]
for p in prompts
]
texts = [
prompt_enhancer_tokenizer.apply_chat_template(
m, tokenize=False, add_generation_prompt=True
)
for m in messages
]
model_inputs = prompt_enhancer_tokenizer(texts, return_tensors="pt").to(
prompt_enhancer_model.device
)
return _generate_and_decode_prompts(
prompt_enhancer_model, prompt_enhancer_tokenizer, model_inputs, max_new_tokens
)
def _generate_i2v_prompt(
image_caption_model,
image_caption_processor,
prompt_enhancer_model,
prompt_enhancer_tokenizer,
prompts: List[str],
first_frames: List[Image.Image],
max_new_tokens: int,
system_prompt: str,
) -> List[str]:
image_captions = _generate_image_captions(
image_caption_model, image_caption_processor, first_frames
)
messages = [
[
{"role": "system", "content": system_prompt},
{"role": "user", "content": f"user_prompt: {p}\nimage_caption: {c}"},
]
for p, c in zip(prompts, image_captions)
]
texts = [
prompt_enhancer_tokenizer.apply_chat_template(
m, tokenize=False, add_generation_prompt=True
)
for m in messages
]
model_inputs = prompt_enhancer_tokenizer(texts, return_tensors="pt").to(
prompt_enhancer_model.device
)
return _generate_and_decode_prompts(
prompt_enhancer_model, prompt_enhancer_tokenizer, model_inputs, max_new_tokens
)
def _generate_image_captions(
image_caption_model,
image_caption_processor,
images: List[Image.Image],
system_prompt: str = "",
) -> List[str]:
image_caption_prompts = [system_prompt] * len(images)
inputs = image_caption_processor(
image_caption_prompts, images, return_tensors="pt"
).to(image_caption_model.device)
with torch.inference_mode():
generated_ids = image_caption_model.generate(
input_ids=inputs["input_ids"],
pixel_values=inputs["pixel_values"],
max_new_tokens=1024,
do_sample=False,
num_beams=3,
)
return image_caption_processor.batch_decode(generated_ids, skip_special_tokens=True)
def _generate_and_decode_prompts(
prompt_enhancer_model, prompt_enhancer_tokenizer, model_inputs, max_new_tokens: int
) -> List[str]:
with torch.inference_mode():
outputs = prompt_enhancer_model.generate(
**model_inputs, max_new_tokens=max_new_tokens
)
generated_ids = [
output_ids[len(input_ids) :]
for input_ids, output_ids in zip(model_inputs.input_ids, outputs)
]
decoded_prompts = prompt_enhancer_tokenizer.batch_decode(
generated_ids, skip_special_tokens=True
)
return decoded_prompts
================================================
FILE: ltx_video/utils/skip_layer_strategy.py
================================================
from enum import Enum, auto
class SkipLayerStrategy(Enum):
AttentionSkip = auto()
AttentionValues = auto()
Residual = auto()
TransformerBlock = auto()
================================================
FILE: ltx_video/utils/torch_utils.py
================================================
import torch
from torch import nn
def append_dims(x: torch.Tensor, target_dims: int) -> torch.Tensor:
"""Appends dimensions to the end of a tensor until it has target_dims dimensions."""
dims_to_append = target_dims - x.ndim
if dims_to_append < 0:
raise ValueError(
f"input has {x.ndim} dims but target_dims is {target_dims}, which is less"
)
elif dims_to_append == 0:
return x
return x[(...,) + (None,) * dims_to_append]
class Identity(nn.Module):
"""A placeholder identity operator that is argument-insensitive."""
def __init__(self, *args, **kwargs) -> None: # pylint: disable=unused-argument
super().__init__()
# pylint: disable=unused-argument
def forward(self, x: torch.Tensor, *args, **kwargs) -> torch.Tensor:
return x
================================================
FILE: pyproject.toml
================================================
[build-system]
requires = ["setuptools>=42", "wheel"]
build-backend = "setuptools.build_meta"
[project]
name = "ltx-video"
version = "0.1.2"
description = "A package for LTX-Video model"
authors = [
{ name = "LTX-Video Team", email = "ltx-video@lightricks.com" }
]
requires-python = ">=3.10"
readme = "README.md"
classifiers = [
"Programming Language :: Python :: 3",
"Operating System :: OS Independent"
]
dependencies = [
"torch>=2.1.0",
"diffusers>=0.28.2",
"transformers>=4.47.2,<4.52.0",
"sentencepiece>=0.1.96",
"huggingface-hub~=0.30",
"einops",
"timm"
]
[project.optional-dependencies]
inference = [
"imageio[ffmpeg]",
"av",
"torchvision"
]
test = [
"pytest",
]
[tool.setuptools.packages.find]
include = ["ltx_video*"]
[tool.setuptools.package-data]
ltx_video = ["configs/*.yaml"]
================================================
FILE: tests/conftest.py
================================================
import json
import pytest
import safetensors.torch
import torch
from ltx_video.models.autoencoders.causal_video_autoencoder import (
CausalVideoAutoencoder,
create_video_autoencoder_demo_config,
PER_CHANNEL_STATISTICS_PREFIX,
)
from ltx_video.models.transformers.transformer3d import Transformer3DModel
def pytest_make_parametrize_id(config, val, argname):
if isinstance(val, str):
return f"{argname}-{val}"
return f"{argname}-{repr(val)}"
@pytest.fixture
def num_latent_channels():
return 16
@pytest.fixture
def video_autoencoder(num_latent_channels):
config = create_video_autoencoder_demo_config(latent_channels=num_latent_channels)
model = CausalVideoAutoencoder.from_config(config)
model.eval().to(torch.bfloat16)
return model
@pytest.fixture
def transformer_config(num_latent_channels):
transformer_config = {
"activation_fn": "gelu-approximate",
"attention_bias": True,
"attention_head_dim": 12,
"attention_type": "default",
"caption_channels": 4096,
"cross_attention_dim": 192,
"double_self_attention": False,
"dropout": 0.0,
"in_channels": num_latent_channels,
"norm_elementwise_affine": False,
"norm_eps": 1e-06,
"norm_num_groups": 32,
"num_attention_heads": 16,
"num_embeds_ada_norm": 1000,
"num_layers": 2,
"num_vector_embeds": None,
"only_cross_attention": False,
"out_channels": num_latent_channels,
"upcast_attention": False,
"use_linear_projection": False,
"qk_norm": "rms_norm",
"standardization_norm": "rms_norm",
"positional_embedding_type": "rope",
"positional_embedding_theta": 10000.0,
"positional_embedding_max_pos": [120, 1, 1],
"timestep_scale_multiplier": 1000,
}
return transformer_config
@pytest.fixture
def synthetic_ckpt_path(
tmp_path, video_autoencoder, num_latent_channels, transformer_config
):
# Create transformer
transformer = Transformer3DModel.from_config(transformer_config)
transformer.to(torch.bfloat16)
# Prepare configs and state dicts
configs = {"transformer": transformer_config, "vae": vars(video_autoencoder.config)}
transformer_sd = transformer.state_dict()
transformer_sd = {
"model.diffusion_model." + key: value for key, value in transformer_sd.items()
}
# Prepare VAE state dict with per-channel statistics
vae_sd = video_autoencoder.state_dict()
vae_sd[f"{PER_CHANNEL_STATISTICS_PREFIX}std-of-means"] = torch.rand(
num_latent_channels,
)
vae_sd[f"{PER_CHANNEL_STATISTICS_PREFIX}mean-of-means"] = torch.rand(
num_latent_channels,
)
vae_sd = {"vae." + key: value for key, value in vae_sd.items()}
out_file_path = f"{tmp_path}/test_ckpt.safetensors"
safetensors.torch.save_file(
{**transformer_sd, **vae_sd},
out_file_path,
metadata={"config": json.dumps(configs)},
)
return out_file_path
================================================
FILE: tests/test_configs.py
================================================
import pytest
from pathlib import Path
from ltx_video.inference import infer, InferenceConfig
CONFIGS_DIR = Path(__file__).parents[1] / "configs"
@pytest.fixture
def prompt():
return "A video of a cat playing with a ball."
# mark as slow to avoid running these tests by default
@pytest.mark.slow
@pytest.mark.parametrize(
"pipeline_config",
[pytest.param(config, id=config.stem) for config in CONFIGS_DIR.glob("*.yaml")],
)
def test_run_config(tmp_path, prompt, pipeline_config):
if "fp8" in pipeline_config.stem:
pytest.skip("Skipping fp8 configs as they require specific hardware support.")
inference_config = InferenceConfig(prompt=prompt)
inference_config.pipeline_config = CONFIGS_DIR / pipeline_config
inference_config.output_path = tmp_path / f"{pipeline_config.stem}"
inference_config.height = 256
inference_config.width = 320
inference_config.num_frames = 33
infer(config=inference_config)
================================================
FILE: tests/test_inference.py
================================================
from dataclasses import asdict
import pytest
import torch
import yaml
from ltx_video.inference import (
create_ltx_video_pipeline,
get_device,
infer,
InferenceConfig,
)
from ltx_video.utils.skip_layer_strategy import SkipLayerStrategy
@pytest.fixture
def input_image_path():
return "tests/utils/woman.jpeg"
@pytest.fixture
def input_video_path():
return "tests/utils/woman.mp4"
def base_inference_config(tmp_path, pipeline_config):
temp_config_path = tmp_path / "config.yaml"
with open(temp_config_path, "w") as f:
yaml.dump(pipeline_config, f)
return InferenceConfig(
seed=42,
height=256,
width=320,
num_frames=49,
frame_rate=25,
prompt="A young woman with wavy, shoulder-length light brown hair stands outdoors on a foggy day. She wears a cozy pink turtleneck sweater, with a serene expression and piercing blue eyes. A wooden fence and a misty, grassy field fade into the background, evoking a calm and introspective mood.",
negative_prompt="worst quality, inconsistent motion, blurry, jittery, distorted",
output_path=tmp_path,
pipeline_config=temp_config_path,
)
@pytest.fixture
def base_pipeline_config(synthetic_ckpt_path):
return {
"num_inference_steps": 1,
"stg_mode": "attention_values",
"skip_block_list": [1],
"precision": "bfloat16",
"decode_timestep": 0.05,
"decode_noise_scale": 0.025,
"checkpoint_path": synthetic_ckpt_path,
"text_encoder_model_name_or_path": "PixArt-alpha/PixArt-XL-2-1024-MS",
"prompt_enhancer_image_caption_model_name_or_path": "MiaoshouAI/Florence-2-large-PromptGen-v2.0",
"prompt_enhancer_llm_model_name_or_path": "unsloth/Llama-3.2-3B-Instruct",
"prompt_enhancement_words_threshold": 120,
"sampler": "LinearQuadratic",
}
@pytest.mark.parametrize(
"conditioning_test_mode",
["unconditional", "first-frame", "first-sequence", "sequence-and-frame"],
ids=lambda x: f"conditioning_test_mode={x}",
)
def test_condition_modes(
tmp_path,
conditioning_test_mode,
input_image_path,
input_video_path,
base_pipeline_config,
):
inference_config = base_inference_config(tmp_path, base_pipeline_config)
if conditioning_test_mode == "unconditional":
pass
elif conditioning_test_mode == "first-frame":
inference_config.conditioning_media_paths = [input_image_path]
inference_config.conditioning_start_frames = [0]
elif conditioning_test_mode == "first-sequence":
inference_config.conditioning_media_paths = [input_video_path]
inference_config.conditioning_start_frames = [0]
elif conditioning_test_mode == "sequence-and-frame":
inference_config.conditioning_media_paths = [input_video_path, input_image_path]
inference_config.conditioning_start_frames = [16, 43]
else:
raise ValueError(f"Unknown conditioning mode: {conditioning_test_mode}")
# Test that the infer function runs without errors
infer(inference_config)
def test_vid2vid(tmp_path, input_video_path, base_pipeline_config):
pipeline_config = base_pipeline_config
pipeline_config["num_inference_steps"] = 3
pipeline_config["skip_initial_inference_steps"] = 1
inference_config = base_inference_config(tmp_path, pipeline_config)
inference_config.num_frames = 25
inference_config.input_media_path = input_video_path
# Test that the infer function runs without errors
infer(inference_config)
def test_pipeline_on_batch(tmp_path, base_pipeline_config):
pipeline_config = base_pipeline_config
inference_config = base_inference_config(tmp_path, pipeline_config)
inference_config.num_frames = 1 # For faster test, we use a single frame
device = get_device()
pipeline = create_ltx_video_pipeline(
ckpt_path=pipeline_config["checkpoint_path"],
device=device,
precision=pipeline_config["precision"],
text_encoder_model_name_or_path=pipeline_config[
"text_encoder_model_name_or_path"
],
enhance_prompt=False,
prompt_enhancer_image_caption_model_name_or_path=pipeline_config[
"prompt_enhancer_image_caption_model_name_or_path"
],
prompt_enhancer_llm_model_name_or_path=pipeline_config[
"prompt_enhancer_llm_model_name_or_path"
],
sampler="LinearQuadratic",
)
first_prompt = "A vintage yellow car drives along a wet mountain road, its rear wheels kicking up a light spray as it moves. The camera follows close behind, capturing the curvature of the road as it winds through rocky cliffs and lush green hills. The sunlight pierces through scattered clouds, reflecting off the car's rain-speckled surface, creating a dynamic, cinematic moment. The scene conveys a sense of freedom and exploration as the car disappears into the distance."
second_prompt = "A woman with blonde hair styled up, wearing a black dress with sequins and pearl earrings, looks down with a sad expression on her face. The camera remains stationary, focused on the woman's face. The lighting is dim, casting soft shadows on her face. The scene appears to be from a movie or TV show."
def get_images(prompts):
generators = [
torch.Generator(device=device).manual_seed(inference_config.seed)
for _ in range(2)
]
torch.manual_seed(inference_config.seed)
params = asdict(inference_config)
params["prompt"] = prompts
pipeline_result = pipeline(
generator=generators,
output_type="pt",
vae_per_channel_normalize=True,
**params,
)
return pipeline_result.images
# Run the pipeline on two different batches of prompts
batch_diff_images = get_images([first_prompt, second_prompt])
batch_same_images = get_images([second_prompt, second_prompt])
# Take the second image from both runs, which should be equal
image2_not_same = batch_diff_images[1, :, 0, :, :]
image2_same = batch_same_images[1, :, 0, :, :]
assert torch.allclose(image2_not_same, image2_same)
def test_prompt_enhancement(tmp_path, base_pipeline_config):
pipeline_config = base_pipeline_config
inference_config = base_inference_config(tmp_path, pipeline_config)
inference_config.num_frames = 1 # For faster test, we use a single frame
device = get_device()
pipeline = create_ltx_video_pipeline(
ckpt_path=pipeline_config["checkpoint_path"],
device=device,
precision=pipeline_config["precision"],
text_encoder_model_name_or_path=pipeline_config[
"text_encoder_model_name_or_path"
],
enhance_prompt=True,
prompt_enhancer_image_caption_model_name_or_path=pipeline_config[
"prompt_enhancer_image_caption_model_name_or_path"
],
prompt_enhancer_llm_model_name_or_path=pipeline_config[
"prompt_enhancer_llm_model_name_or_path"
],
sampler="LinearQuadratic",
)
# Mock the pipeline's _encode_prompt method to verify the prompt being used
original_encode_prompt = pipeline.encode_prompt
def mock_encode_prompt(prompt, *args, **kwargs):
prompts_used.append(prompt[0] if isinstance(prompt, list) else prompt)
return original_encode_prompt(prompt, *args, **kwargs)
pipeline.encode_prompt = mock_encode_prompt
original_prompt = "A cat sitting on a windowsill"
inference_config.prompt = original_prompt
def run_pipeline(enhance_prompt):
params = asdict(inference_config)
pipeline(
enhance_prompt=enhance_prompt,
**params,
skip_layer_strategy=SkipLayerStrategy.AttentionValues,
vae_per_channel_normalize=True,
output_type="pt",
)
assert (
len(prompts_used) > 0
), f"No prompts were used in the pipeline run with enhance_prompt={enhance_prompt}"
if enhance_prompt:
# Verify that the enhanced prompt was used
assert (
prompts_used[0] != original_prompt
), f"Expected enhanced prompt to be different from original prompt, but got: {original_prompt}"
else:
# Verify that the original prompt was used
assert (
prompts_used[0] == original_prompt
), f"Expected original prompt to be used, but got: {prompts_used[0]}"
# Run pipeline with prompt enhancement enabled
prompts_used = []
run_pipeline(enhance_prompt=True)
# Run pipeline with prompt enhancement disabled
prompts_used = []
run_pipeline(enhance_prompt=False)
================================================
FILE: tests/test_scheduler.py
================================================
import pytest
import torch
from ltx_video.schedulers.rf import RectifiedFlowScheduler
def init_latents_and_scheduler(sampler):
batch_size, n_tokens, n_channels = 2, 4096, 128
num_steps = 20
scheduler = RectifiedFlowScheduler(
sampler=("Uniform" if sampler.lower() == "uniform" else "LinearQuadratic")
)
latents = torch.randn(size=(batch_size, n_tokens, n_channels))
scheduler.set_timesteps(num_inference_steps=num_steps, samples_shape=latents.shape)
return scheduler, latents
@pytest.mark.parametrize("sampler", ["LinearQuadratic", "Uniform"])
def test_scheduler_default_behavior(sampler):
"""
Test the case of a single timestep from the list of timesteps.
"""
scheduler, latents = init_latents_and_scheduler(sampler)
for i, t in enumerate(scheduler.timesteps):
noise_pred = torch.randn_like(latents)
denoised_latents = scheduler.step(
noise_pred,
t,
latents,
return_dict=False,
)[0]
# Verify the denoising
next_t = scheduler.timesteps[i + 1] if i < len(scheduler.timesteps) - 1 else 0.0
dt = t - next_t
expected_denoised_latents = latents - dt * noise_pred
assert torch.allclose(denoised_latents, expected_denoised_latents, atol=1e-06)
@pytest.mark.parametrize("sampler", ["LinearQuadratic", "Uniform"])
def test_scheduler_per_token(sampler):
"""
Test the case of a timestep per token (from the list of timesteps).
Some tokens are set with timestep of 0.
"""
scheduler, latents = init_latents_and_scheduler(sampler)
batch_size, n_tokens = latents.shape[:2]
for i, t in enumerate(scheduler.timesteps):
timesteps = torch.full((batch_size, n_tokens), t)
timesteps[:, 0] = 0.0
noise_pred = torch.randn_like(latents)
denoised_latents = scheduler.step(
noise_pred,
timesteps,
latents,
return_dict=False,
)[0]
# Verify the denoising
next_t = scheduler.timesteps[i + 1] if i < len(scheduler.timesteps) - 1 else 0.0
next_timesteps = torch.full((batch_size, n_tokens), next_t)
dt = timesteps - next_timesteps
expected_denoised_latents = latents - dt.unsqueeze(-1) * noise_pred
assert torch.allclose(
denoised_latents[:, 1:], expected_denoised_latents[:, 1:], atol=1e-06
)
assert torch.allclose(denoised_latents[:, 0], latents[:, 0], atol=1e-06)
@pytest.mark.parametrize("sampler", ["LinearQuadratic", "Uniform"])
def test_scheduler_t_not_in_list(sampler):
"""
Test the case of a timestep per token NOT from the list of timesteps.
"""
scheduler, latents = init_latents_and_scheduler(sampler)
batch_size, n_tokens = latents.shape[:2]
for i in range(len(scheduler.timesteps)):
if i < len(scheduler.timesteps) - 1:
t = (scheduler.timesteps[i] + scheduler.timesteps[i + 1]) / 2
else:
t = scheduler.timesteps[i] / 2
timesteps = torch.full((batch_size, n_tokens), t)
noise_pred = torch.randn_like(latents)
denoised_latents = scheduler.step(
noise_pred,
timesteps,
latents,
return_dict=False,
)[0]
# Verify the denoising
next_t = scheduler.timesteps[i + 1] if i < len(scheduler.timesteps) - 1 else 0.0
next_timesteps = torch.full((batch_size, n_tokens), next_t)
dt = timesteps - next_timesteps
expected_denoised_latents = latents - dt.unsqueeze(-1) * noise_pred
assert torch.allclose(denoised_latents, expected_denoised_latents, atol=1e-06)
================================================
FILE: tests/test_vae.py
================================================
import pytest
import torch
from ltx_video.models.autoencoders.causal_video_autoencoder import (
CausalVideoAutoencoder,
)
def test_encode_decode_shape(video_autoencoder, num_latent_channels):
spatial_factor = video_autoencoder.spatial_downscale_factor
temporal_factor = video_autoencoder.temporal_downscale_factor
input_videos = torch.randn(2, 3, 17, 64, 64, dtype=torch.bfloat16)
# Encode
latent = video_autoencoder.encode(input_videos).latent_dist.mode()
expected_shape = (
input_videos.shape[0],
num_latent_channels,
(input_videos.shape[2] + 7) // temporal_factor,
input_videos.shape[3] // spatial_factor,
input_videos.shape[4] // spatial_factor,
)
assert latent.shape == expected_shape
# Decode
timestep = torch.ones(input_videos.shape[0]) * 0.1
reconstructed_videos = video_autoencoder.decode(
latent, target_shape=input_videos.shape, timestep=timestep
).sample
assert input_videos.shape == reconstructed_videos.shape
def test_temporal_causality(video_autoencoder):
# validate temporal causality in encoder
input_videos = torch.randn(2, 3, 17, 64, 64, dtype=torch.bfloat16)
latent = video_autoencoder.encode(input_videos).latent_dist.mode()
# Check that encoding a single frame matches the corresponding slice in the full latent
input_image = input_videos[:, :, :1, :, :]
image_latent = video_autoencoder.encode(input_image).latent_dist.mode()
assert torch.allclose(image_latent, latent[:, :, :1, :, :], atol=1e-6)
# Check that encoding a sequence of frames matches the corresponding slice in the full latent
input_sequence = input_videos[:, :, :9, :, :]
sequence_latent = video_autoencoder.encode(input_sequence).latent_dist.mode()
assert torch.allclose(sequence_latent, latent[:, :, :2, :, :], atol=1e-6)
@pytest.mark.parametrize(
"layer_name,expected_temporal_factor,expected_spatial_factor",
[
("compress_space_res", 1, 2),
("compress_space", 1, 2),
("compress_time_res", 2, 1),
("compress_time", 2, 1),
("compress_all_res", 2, 2),
("compress_all", 2, 2),
],
)
def test_downscale_factors(
num_latent_channels, layer_name, expected_temporal_factor, expected_spatial_factor
):
patch_size = 4
encoder_blocks = [
(layer_name, {"multiplier": 2}),
]
decoder_blocks = [
("compress_all", {"residual": True, "multiplier": 2}),
]
config = {
"_class_name": "CausalVideoAutoencoder",
"dims": 3,
"encoder_blocks": encoder_blocks,
"decoder_blocks": decoder_blocks,
"latent_channels": num_latent_channels,
"norm_layer": "pixel_norm",
"patch_size": patch_size,
"latent_log_var": "uniform",
"use_quant_conv": False,
"causal_decoder": False,
"timestep_conditioning": True,
"spatial_padding_mode": "replicate",
}
model = CausalVideoAutoencoder.from_config(config)
assert model.temporal_downscale_factor == expected_temporal_factor
assert model.spatial_downscale_factor == expected_spatial_factor * patch_size
================================================
FILE: tests/utils/.gitattributes
================================================
*.mp4 filter=lfs diff=lfs merge=lfs -text