kurttu4

set CUDA_PATH=venv\Lib\site-packages\torch\lib or yours in .bat & check .dll in folder & python -c "import torch; print(torch.cuda.is_available())" or Delete venv\Lib\site-packages\torch folder & reinstall

u have in venv\Lib\site-packages\torch\lib folder .dll 

What about Humo? The decoder and lora aren't loading there. All the lores are downloaded from links in the workflow. ``` missing audio encoder: ['feature_extractor.mel_spectrogram.spectrogram.window', 'feature_extractor.mel_spectrogram.mel_scale.fb'] unexpected audio encoder: ['decoder.embed_positions.weight',...

> here is a little script to check for keys in dict if model has BLF (black forest lab) or diffusers format: > kijai/ComfyUI-WanVideoWrapper - works perfectly with the same...

@Eraxorice So, what's wrong with the generation? My lips weren't moving in sync with the sound. I thought maybe it was due to errors, but now the same errors are...

@Eraxorice I created a node with additional settings for the influence of text and audio on generation ,Maybe it will help. https://github.com/kurttu4/wanhumo_custom_sampler