fast-stable-diffusion
fast-stable-diffusion copied to clipboard
Add SD 1.4 as an option
Could we get the version as an option? It doesn't have to be a UI item, can just be a variable for us to change. Currently it's not too hard to do, but the string is hardcoded to too many places.
I'll add it soon
It'd be really nice.
Actually, what's the difference between the 1.4 and 1.5 dreambooth notebook? Is it just the model downloaded from hugging face being different?
Has there been any other change to the actual training process? (apart from changes to the learning rate and text encoder)
No, if you change the model from Huggingface everything else works fine.
No, if you change the model from Huggingface everything else works fine.
Good to know. Thanks
was thinking of doing a model merge to see what happens because im currently hating the pruned 1.5 ema only model 75% of the time. it really follows the prompt well which i like but damn it is atrociously ugly for what im doing. im probably doing something wrong tho