dreambooth-template
dreambooth-template copied to clipboard
replicate
A template repo for training and publishing your own custom Stable Diffusion model using https://replicate.com/replicate/dreambooth
NOTE: we are currently NOT using `COG_WEIGHTS` as we either double download or are stuck with a slow download This template model uses weights specified by URL set in`DREAMBOOTH_WEIGHTS`
I know there's some recent work on Replicate to improve SD speeds almost 2x - how difficult would it be to incorporate that magic into this template?
When the [`image`](https://github.com/replicate/dreambooth-template/blob/115ba1d44965dd2fe38f70d10835e94b43daf783/predict.py#L91-L94) input is supplied, the model attempts to create an output image of the same dimensions. Users should ideally be providing square images of smaller dimensions like 512x512,...
Users have asked for support for longer prompts This might be the answer https://github.com/huggingface/diffusers/tree/main/examples/community#stable-diffusion-mega
I've seen this feature requested multiple times in the Discord channel and could only find one answer from Jesse. I assume it's not possible currently, so I'm adding this request...
To fully customize prompts, there is a "hack" that popular versions of stable diffusion use to allow you to increase/decrease the strength of each token using parens/square brakets [lpw_stable_diffusion.py](https://github.com/huggingface/diffusers/blob/main/examples/community/lpw_stable_diffusion.py) is...
It would be nice if you could mask/inpaint with your dreambooths. while stable diffusion 1.5 / 2.x has moved to having specialized models/weights for inpainting, as @andreasjansson says: > yes,...
I guess downloading weights for safety checker requires python packages diffusers and torch - should this be documented?