Support for HunyuanVideo Loras created via Replicate tool?
- Dominant language
- Python
- Stars
- 133k
- Forks
- 15.7k
- Avg merge
- 1d 7h
- Merged PRs (30d)
- 158
Description
### Feature Idea
I've been using comfyui via SwarmUI. Both tools are up to date.
I've been finding a lot of HunyuanVideo Loras on huggingface.co that won't load.
e.g. https://huggingface.co/fofr/hunyuan-take-on-me/tree/main
`
11:24:45.897 [Debug] [GridGenerator] Pre-prepping 8/982 ... Set: Text2Video Frames=1, LoRA Weights=0.7, LoRAs=_hunyuan_loras_offload/2025-01-11/hunyuan-take-on-me, Seed=1, file 1/hunyuantakeonme/0_7/1
11:24:45.924 [Debug] [ComfyUI-0/STDERR] lora key not loaded: lora_unet_double_blocks_12_img_mlp_fc2.lora_up.weight
11:24:45.929 [Debug] [ComfyUI-0/STDERR] lora key not loaded: lora_unet_double_blocks_12_img_mlp_fc1.lora_up.weight
11:24:45.934 [Debug] [ComfyUI-0/STDERR] lora key not loaded: lora_unet_single_blocks_30_modulation_linear.lora_up.weight
11:24:45.939 [Debug] [ComfyUI-0/STDERR] lora key not loaded: lora_unet_double_blocks_6_txt_mlp_fc2.alpha
11:24:45.945 [Debug] [ComfyUI-0/STDERR] lora key not loaded: lora_unet_double_blocks_17_txt_mlp_fc2.alpha
11:24:45.951 [Debug] [ComfyUI-0/STDERR] lora key not loaded: lora_unet_double_blocks_2_img_mlp_fc2.alpha
11:24:45.957 [Debug] [ComfyUI-0/STDERR] lora key not loaded: lora_unet_double_blocks_1_img_mlp_fc2.lora_up.weight
11:24:45.964 [Debug] [ComfyUI-0/STDERR] lora key not loaded: lora_unet_single_blocks_11_modulation_linear.lora_up.weight
...
`
Apparently the Loras are being trained using this tool:
https://replicate.com/zsxkib/hunyuan-video-lora/train
Any chances this format could be incorporated into comfy?
### Existing Solutions
_No response_
### Other
_No response_
Contributor guide
Assessment
This issue has not been assessed yet.