AI GitHub
LTX-Video for prompt and image-conditioned clips
Run documented generation pipelines for text and visual conditioning.
Key Takeaways
- The project packages models and inference configurations for producing video clips.
- Choose a documented conditioning mode and model variant.
- Test a short motion prompt and an image-conditioned shot.
- Supported dimensions and performance depend on the checkpoint and pipeline.
How it works
The project packages models and inference configurations for producing video clips. The selected workflow determines whether a source frame or text provides the conditioning.
What you need
Choose a documented conditioning mode and model variant. Set the clip dimensions and target duration before choosing the accelerator configuration.
What to try first
Test a short motion prompt and an image-conditioned shot. Inspect how motion, subject identity and detail change across the clip rather than choosing a single preview frame.
Use an initial image with a clear background and a moving subject. It makes camera motion, object motion and drifting background features easier to distinguish.
Where to look for failures
Supported dimensions and performance depend on the checkpoint and pipeline. Budget for multiple attempts when continuity matters.
Related Franklin pages
Compare complete clips through playback and frame inspection. Keep prompt, duration and dimensions as similar as supported, and inspect whether subject and camera motion match the task.
- Use Tencent's reference implementation and checkpoints for generating clips from text. HunyuanVideo for experimenting with text-to-video
- Explore the official CogVideoX code and workflows for supported video tasks. CogVideo for experimenting with video diffusion models
For video assembled from authored HTML and motion graphics, see Franklin’s HyperFrames profile. HyperFrames
Project documentation
Try it for your workflow
Explore the project’s documentation for setup and usage, or save this page to a collection for later.
Read the project documentation ↗