AI GitHub
Wan 2.1 for running video generation checkpoints
Experiment with the official Wan text-to-video and image-to-video workflows.
Key Takeaways
- The repository provides inference instructions and model variants for several generation tasks.
- Select the checkpoint for text or image conditioning and use the repository's matching pipeline.
- Generate one short shot and compare results across two seeds.
- Video length and resolution raise memory requirements.
How it works
The repository provides inference instructions and model variants for several generation tasks. Checkpoint and configuration choices determine how conditioning and output sizes are handled.
What you need
Select the checkpoint for text or image conditioning and use the repository's matching pipeline. Keep original input frames and generation settings together for inspection.
What to try first
Generate one short shot and compare results across two seeds. In an image-conditioned test, check whether objects and framing persist through the clip.
Measure end-to-end runtime including model load and decoding. Check the generated frame count and playback rate before comparing duration with another run.
Where to look for failures
Video length and resolution raise memory requirements. Check checkpoint terms and avoid extrapolating performance from a still frame.
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.
- Run documented generation pipelines for text and visual conditioning. LTX-Video for prompt and image-conditioned clips
- Use Tencent's reference implementation and checkpoints for generating clips from text. HunyuanVideo for experimenting with text-to-video
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 ↗