error in google collab
#17
by
TheBestButNot
- opened
+1
If the revision is removed it works, presumably because there is an issue with the fp16 branch. Now I'm using the main branch instead of fp16. What's the difference between those branches?
fp16 is float16 while main branch is float 32, I think that's the difference
I've only used fp16 and on main branch I see CUDA running out of memory for larger than 512x512 images on T4. So I don't know if this is branch related or all-branches-update related, but interested to test anyway. I just want to make larger images.
im terrible at using git, please try it again and let me know if it doesnt work
should work now
hakurei
changed discussion status to
closed