Robotic Manipulation on DynaGrasp-32
81.88Success RateX-VLA + DynaWM
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| X-VLA + DynaWMCheckpoint=X-VLA, Tuning Setting=Fine-tuned, DynaWM Integration=True2026.07 | 81.88 | 45.31 | 32 | 0 | 0 | |
| X-VLA + DynaWMCheckpoint=X-VLA, Tuning Setting=Coarse-tuned, DynaWM Integration=True2026.07 | 81.75 | 44.06 | 32 | 0 | 0 | |
| SmolVLA + DynaWMCheckpoint=SmolVLA, Tuning Setting=Coarse-tuned, DynaWM Integration=True2026.07 | 75.75 | 35.13 | 32 | 0 | 0 | |
| π0.5 + DynaWMCheckpoint=π0.5, Tuning Setting=Fine-tuned, DynaWM Integration=True2026.07 | 75.56 | 10.94 | 21 | 5 | 6 | |
| π0 + DynaWMCheckpoint=π0, Tuning Setting=Coarse-tuned, DynaWM Integration=True2026.07 | 75.56 | 35.69 | 31 | 1 | 0 | |
| π0.5 + DynaWMCheckpoint=π0.5, Tuning Setting=Coarse-tuned, DynaWM Integration=True2026.07 | 74.56 | 26.13 | 31 | 0 | 1 | |
| π0 + DynaWMCheckpoint=π0, Tuning Setting=Fine-tuned, DynaWM Integration=True2026.07 | 74.44 | 1.88 | 14 | 4 | 14 | |
| π0Checkpoint=π0, Tuning Setting=Fine-tuned, DynaWM Integration=False2026.07 | 72.56 | — | — | — | — | |
| SmolVLA + DynaWMCheckpoint=SmolVLA, Tuning Setting=Fine-tuned, DynaWM Integration=True2026.07 | 69.69 | 7.19 | 23 | 1 | 8 | |
| π0.5Checkpoint=π0.5, Tuning Setting=Fine-tuned, DynaWM Integration=False2026.07 | 64.63 | — | — | — | — | |
| SmolVLACheckpoint=SmolVLA, Tuning Setting=Fine-tuned, DynaWM Integration=False2026.07 | 62.5 | — | — | — | — | |
| π0.5Checkpoint=π0.5, Tuning Setting=Coarse-tuned, DynaWM Integration=False2026.07 | 48.44 | — | — | — | — | |
| SmolVLACheckpoint=SmolVLA, Tuning Setting=Coarse-tuned, DynaWM Integration=False2026.07 | 40.63 | — | — | — | — | |
| π0Checkpoint=π0, Tuning Setting=Coarse-tuned, DynaWM Integration=False2026.07 | 39.88 | — | — | — | — | |
| X-VLACheckpoint=X-VLA, Tuning Setting=Coarse-tuned, DynaWM Integration=False2026.07 | 37.69 | — | — | — | — | |
| X-VLACheckpoint=X-VLA, Tuning Setting=Fine-tuned, DynaWM Integration=False2026.07 | 36.56 | — | — | — | — |