Fix gitignore.
Add epoch to db_config.
Compile lora checkpoints when training.
Remove unused messages and files.
Detect 512 V2 models and save the appropriate config file.
Fix global step counter.
Cleanup startup script to not freak people out.
Use pypi diffusers version, bump transformers version.
Ensure requirements are installed.
Use Shiv's repo, implement multiple concept training, etc.
Split dreambooth and train_dreambooth to make future updates easier.
Rename train_config.py
UI Work (load from save needs fixing yet)