Will Berman
Will Berman
> Let me know if you need help with anything :-) Some help with the prior transformer would be great! cc @patil-suraj I think you said you might be able...
@nbanyan the cli arg `--checkpoints_total_limit` will now limit the total number of checkpoints saved. Sorry for the delay here :)
> * By default, we should use the highest precision, least optimized training options > * We allow all kinds of optimized training options (mixed precision, xformers, ....) and let...
Ok, I think I follow everything tl;dr: training/fine tuning shouldn't be done with fp16 weights[^1], fp16 inputs are ok with amp + gradient scaling. fp16 weights throw an error when...
Since we merged the guard on loading low precision weights, going to close this issue :)
Hey folks! We're trying to encourage the forum for open ended discussion :) Might be good to make a thread there for future dreambooth inpainting discussion https://discuss.huggingface.co/c/discussion-related-to-httpsgithubcomhuggingfacediffusers/63
Looks interesting! Don't have immediate time but will add to my backlog
@patil-suraj yep still standing, haven't had time to look!
Nice! All comments are small things. Looks basically good to go
@patrickvonplaten yeah I think it happened when I merged the other PR