BREAKING
Nanjing U & Tencent release SteadyDancer
0B
parameters
0
released
I2V replaces the R2V approach
R2Vold
Spatial mismatches
Identity drift
I2VSteadyDancer
First-frame anchor
Preserves identity
How SteadyDancer works
1Anchor first frame
2Track pose
3Generate video
Multiple distribution formats
Standard14B
Hugging Face weights
QuantizedGGUF
Lightweight inference
For China14B
ModelScope
ComfyUI support still maturing
AI NEWS BLITZ
Nanjing University and Tencent have unveiled a new human animation model called SteadyDancer.