vadim epstein

not really; i expect some way to run a command (used in batches) triggering comfy to perform the actions, rather than externally reproducing the pipeline (which may and probably will...

@LittlelittleZ check https://github.com/clovaai/stargan-v2/blob/master/README.md#animal-faces-hq-dataset-afhq

number of layers depends on the model resolution. (1,18,512) is for 1024x1024, (1,12,512) is for 128x128. you should either take bigger model, or manually change that hard-coded shape.

thanks for a prompt response! here is few examples: https://github.com/Lightricks/LongAnimateDiff https://huggingface.co/guoyww/animatediff/tree/main [see v3 versions]

+1, just yesterday tried fresh release - from_single_file not implemented yet for motion adapter

> Hi @eps696 have you tried using the AnimateDiff conversion scripts with those checkpoints? no, missed it somehow, thanks for a hint!