Being-M0.5: A Real-Time Controllable Vision-Language-Motion Model

Benchmark Model Rank Results
motion-synthesis-on-humanml3dBeing-M0.5-PRQ 4–FID: 0.056R Precision Top3: 0.821
motion-synthesis-on-humanml3dBeing-M0.5-VQ 1–FID: 0.141R Precision Top3: 0.815