[DOCS] LTX Video Generation Docs #3232
Conversation
There was a problem hiding this comment.
Pull request overview
This PR introduces comprehensive documentation for LTX Video Generation, explaining how to use the OpenVINO GenAI library to generate videos from text prompts using the Lightricks/LTX-Video model.
Changes:
- Added documentation for LTX-Video pipeline with technical overview and workflow diagram
- Provided code examples for both Python and C++ implementations with CPU/GPU device support
- Documented video generation configuration parameters and usage options
Reviewed changes
Copilot reviewed 6 out of 6 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
| site/static/img/ltx-pipeline.png | Added pipeline workflow diagram using Git LFS |
| site/docs/use-cases/video-generation/index.mdx | Main documentation page with LTX-Video overview and model conversion instructions |
| site/docs/use-cases/video-generation/_sections/_usage_options/index.mdx | Configuration parameters and usage examples for video generation |
| site/docs/use-cases/video-generation/_sections/_run_model/index.mdx | Device-specific implementation instructions for CPU and GPU |
| site/docs/use-cases/video-generation/_sections/_run_model/_text2video_python.mdx | Python code example with video saving utility |
| site/docs/use-cases/video-generation/_sections/_run_model/_text2video_cpp.mdx | C++ code example for text-to-video generation |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
| # highlight-end | ||
| ).video | ||
|
|
||
| save_video("genai_video.avi", video, fps) |
There was a problem hiding this comment.
Variable name inconsistency: the function parameter is video_tensor (line 24) but it's being called with video (line 37). These should match for clarity.
| save_video("genai_video.avi", video, fps) | |
| save_video("genai_video.avi", video_tensor, fps) |
There was a problem hiding this comment.
updated tensor name to video:
| save_video("genai_video.avi", video, fps) | |
| video = pipe.generate(...) |
site/docs/use-cases/video-generation/_sections/_run_model/_text2video_python.mdx
Outdated
Show resolved
Hide resolved
| writer.release() | ||
| print(f"Wrote {output_path} ({num_frames} frames, {width}x{height} @ {fps} fps)") | ||
|
|
||
|
|
There was a problem hiding this comment.
Undefined variable model_path is used without being defined. This should be defined earlier or passed as a parameter to make the code example runnable.
| model_path = "path/to/text2video/model" |
site/docs/use-cases/video-generation/_sections/_run_model/index.mdx
Outdated
Show resolved
Hide resolved
site/docs/use-cases/video-generation/_sections/_run_model/_text2video_python.mdx
Outdated
Show resolved
Hide resolved
| # highlight-end | ||
| ).video | ||
|
|
||
| save_video("genai_video.avi", video, fps) |
There was a problem hiding this comment.
Please also add Video Generation use case to the home page and reuse _text2video_cpp.mdx & _text2video_python.mdx snippets.
Here is the UseCasesSection - https://github.com/likholat/openvino.genai/blob/video_gen_docs/site/src/pages/_sections/UseCasesSection/index.tsx#L21 , just add new component VideoGeneration similary to ImageGeneration
2 participants
https://likholat.github.io/openvino.genai/