abstract_liquid

Motion Lora for AnimateDiff for SD 1.5

Use with v3_sd15_mm.ckpt for v1.0
and
AnimateLCM_sd15_t2v for v1.0_LCM

This lora is pretty over fit so I would recommend starting between 0.5 - 0.6 or otherwise you will see the training data for the most part.

Will do more testing on safetensors at different training steps.