Human Image Animation on TikTok (sequences 335 to 340)
5.9FID-VIDImplicit Preference Alignment
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Implicit Preference AlignmentPretrained model=VACE-14B, Backbone=Wan foundational model, Training=LoRA rank 1282026.05 | 5.9 | 255 | 84.1 | 23.8 | |
| Wan-AnimateBackbone=Wan foundational model2026.05 | 8.6 | 316 | 79.9 | 20.5 | |
| MimicMotionBackbone=Stable Video Diffusion2026.05 | 9.3 | 594 | 79.5 | 20.1 | |
| Wan2.2-Fun-A14B-ControlBackbone=Wan foundational model2026.05 | 9.8 | 360 | 77.3 | 20.7 | |
| UniAnimate-DiTBackbone=Wan foundational model2026.05 | 11.8 | 350 | 76.8 | 20.4 | |
| Moore-AnimateAnyoneBackbone=Stable Diffusion v1.52026.05 | 12.4 | 728 | 75.8 | 18.7 | |
| MagicPoseBackbone=Stable Diffusion v1.52026.05 | 13.3 | 916 | 77.6 | 18.8 | |
| VACEBackbone=Wan foundational model2026.05 | 13.4 | 427 | 77.7 | 20.2 | |
| MuseVBackbone=Stable Diffusion v1.52026.05 | 14.6 | 754 | 76.6 | 17.6 | |
| MagicAnimateBackbone=Stable Diffusion v1.52026.05 | 16.2 | 848 | 74 | 17.5 |