Open GavChap opened 1 month ago
Please add better support for training.
You can use the cache function of diffusers pipeline: https://huggingface.co/docs/diffusers/main/en/api/diffusion_pipeline#diffusers.DiffusionPipeline.from_pretrained.cache_dir
Hi, is there a way to lora train without having to download T5 and the base model from huggingface again? I don't need 3-4 copies of T5 on my computer