Instructions to use Kijai/WanVideo_comfy_fp8_scaled with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusion Single File
How to use Kijai/WanVideo_comfy_fp8_scaled with Diffusion Single File:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
e5m2 Version for Animate V2?
Hi I noticed there wasn't a e5m2 version of the Wan Animate V2. I have a Fp16 version. Would you like me to contribute it to this repo for older GPUs or is this a work in progress?
Just use Native and place in same model. This is a WanVideoWrapper Issue.
Just use Native and place in same model. This is a WanVideoWrapper Issue.
If you mean e4m3fn not working on 3000 series and older, it's not wrapper issue, it's torch.compile issue, will happen in native too when using compile.
For whatever reason. It works in native, but not in the WanVideoWrapper. Even though its not supposed to be like that.
For whatever reason. It works in native, but not in the WanVideoWrapper. Even though its not supposed to be like that.
Torch.compile with e4m3fn on 3000 series card works in native now? It never used to before at least.
