-
-
Notifications
You must be signed in to change notification settings - Fork 60
[FEATURE REQUEST] add download of the ckpt from hugging face git LFS instead of sending users there directly #4
Comments
I find that this does the trick: |
This needs a checkbox for sure, for folks who use colab alot, they already have the model in their drive, this could just mean un-nessesary download time. Care to open a PR with how you think this should work? I'd also comment that I think it's a good idea to still save this model if it doesn't exist in users Gdrive, as well as outputs, so when the colab closes on them, they will still have the results |
Could be super cool if we... check if the model exists already in the specified model folder (either in drive or locally) and then if it doesn't.. ask them for their hugging face token and download it for them |
I just use this, it downloads in about 30 seconds:
Then move to folder:
|
I won't add this to the repo as it would violate the hugging face TOC I think |
sorry haha stupid question |
That's all good, and it works! I did this when I walked some of my friends through this. |
Possible to have the notebook download the model for people with slow upload speeds?
I ended up using this: https://gist.github.com/davo/4b374f41be657f59bfb60d1ee61d2616 to save myself a lot of time
Originally posted by @pinilpypinilpy in Sygil-Dev/stable-diffusion#108 (comment)
The text was updated successfully, but these errors were encountered: