kijai / ComfyUI-CogVideoXWrapper

982 stars 59 forks source link

Error no file named diffusion_pytorch_model.bin ... #269

Closed GeffersAtWork closed 18 hours ago

GeffersAtWork commented 1 day ago

Using the example workflow for 1.5 I2V from github.

Manually downloaded all the files I could find on hugging face, all boxes are good, but that message pops up before anything happens.

"Error no file named diffusion_pytorch_model.bin found in directory C:\ComfyUI_windows_portable\ComfyUI\models\CogVideo\CogVideoX-5b-1.5."

KrakeyMTL commented 1 day ago

It didn't fully download all the files. I get this sometimes when I manually put them in - comfyui likes to..do things by itself. annoying.

I got around this by using the built in node downloader that Kijai set into the wrapper.

When comfyui "fetches" itself it works. odd I know but I get it randomly as well. I think this is more a comfy thing

kijai commented 1 day ago

Huggingface downloader occasionally has these issues especially with large files, I've started to setup alternative cleaner way of getting the models which can be loaded with these new nodes:

image

image

Not all models are converted yet, but I have the 1.5 available here:

https://huggingface.co/Kijai/CogVideoX-comfy/tree/main

The models go to the normal ComfyUI/models/diffusion_models folder and VAE to ComfyUI/models/vae

GeffersAtWork commented 1 day ago

Thanks, much appreciated, downloading now.

GeffersAtWork commented 1 day ago

Well, downloaded those files, but it doesn't help. Still getting that specific file missing, and it's true, it doesn't exist. "diffusion_pytorch_model.bin" - where is it? There is a version of it in the previous version folder. CogvideoXWrapper is highlighted, suggesting it's a file missing related to the CogVideo Downloader. Diffusion

kijai commented 22 hours ago

Well, downloaded those files, but it doesn't help. Still getting that specific file missing, and it's true, it doesn't exist. "diffusion_pytorch_model.bin" - where is it? There is a version of it in the previous version folder. CogvideoXWrapper is highlighted, suggesting it's a file missing related to the CogVideo Downloader. Diffusion

Use the another node I showed above, not the download node. And the models go to the normal comfyui diffusion_models folder.

GeffersAtWork commented 18 hours ago

Yes, I had put them there too... Thanks okay looks better now - I was using the example workflow from github in the examples folder, but it evidently isn't updated so was using the old nodes. cheers :-)

GeffersAtWork commented 18 hours ago

Thanks all. Example workflow needed the nodes changing to the new ones.