Hershel Don
Hershel Don
> So, I edited my UI settings to adjust the values (allow for CFG scale of 0.15). Output images were...not great? > > Looking at the repo of the fellow...
These two commits make unet and text encoder lora resumable by providing the path to the respective .pt files in 2 new arguments to train_lora_dreambooth.py. Lora.py is modified so the...
Somewhere after 750 the speed went back up. Not sure why, no different activity on the computer.
I think it was a background program causing the slowdown
Not a background program, observed the same pattern of speedup and slowdown between checkpoints with no background processes running.
> I also gave up on training session last night because I kept ending up at 7s or 3s per it when I usually train at ~1.6s/it. > > Is...
resolved in new updates! thanks!
What are you trying to do when you're running out of memory?
https://github.com/cloneofsimo/lora It is a smaller approximation of dreambooth. Doesn't work properly with gradient checkpointing for now, so don't waste time training with that on.
> I have 1080TI and I'm able to train v1.5 512 with or without lora currently