AnimateDiff
AnimateDiff copied to clipboard
how to train a v2 model?
Hi!Thank you for your excellent work. I tried to finetune the "mm_sd_v15_v2.ckpt" model by adding my own datasets. It seems that the checkpoint I got doesn't work well. I think it is because the trainning code is used for version1. There are some keys missing or unexpected.
Could you please provide the code for training v2 model? Or is there any method to adapt current code?