ai-toolkit
ai-toolkit copied to clipboard
Can we speedup the training with Hyper lora?
https://huggingface.co/ByteDance/Hyper-SD
Can we make the origial FLUX model Hyper version, than train on it?