Verified: a year ago
PickleTensor
Type | |
Stats | 11,005 |
Reviews | |
Published | Oct 7, 2023 |
Base Model | |
Training | Steps: 111,560 |
Usage Tips | Clip Skip: 2 |
Hash | AutoV2 D683556A3B |
Fine-tune of AnimateDiff mm_sd_v15_v2.ckpt to improve coherency with moving objects
Should generally have better coherence, but can be worse for some cases
Tested with ComfyUI AnimateDiff
Stable diffusion model used in examples: DreamShaper8
Recommended to use the negatives embeddings as well: BadDream, UnrealisticDream
Examples have been interpolated from 8 fps to 16 fps