SFM: Taming State Space Models for Text-to-Motion via Spatial-Frequency Modeling

Benchmark Model Rank Results
motion-synthesis-on-humanml3dSFM–R Precision Top3: 0.845