Datasets:
How to use with "Last Ben" dreambooth
Hi,
How exactly do I use these with Last Ben's db, what is the workflow?
TYVM
So, unfortunately it seems like TheLastBen's Github notebooks have some hardcoding for the regularization images being used. Luckily it should be pretty simple to fix.
For this notebook for example, it looks like you need to point the CLASS_DIR variable to the path to your custom regularization images and ensure that With_Prior_Preservation is set to true:
https://github.com/TheLastBen/fast-stable-diffusion/blob/main/fast-DreamBooth.ipynb
You also need to make sure that the class word / prompt matches the dataset one.
You also need to make sure that the class word / prompt matches the dataset one.
How do I do that, exactly?
If the session name is < keyword1 >_< keywordn >, what do I need to name the custom regularization images to? Or their container folder, idk?
@SAC020 I would suggest asking TheLastBen on how to use regularization images, here: https://github.com/TheLastBen/fast-stable-diffusion/issues
He's going to be able to be able to help you with is code far better than I can.