Stacking on Real-robot manipulation environment
81.5Success RateSmolVLA
Evaluation results
| Method | Links | |
|---|---|---|
| SmolVLATraining=Multi-task2026.03 | 81.5 | |
| One-Step VLATraining=Multi-task2026.03 | 64 | |
| Diffusion PolicyTraining=Single-task2026.03 | 46 |
| Method | Links | |
|---|---|---|
| SmolVLATraining=Multi-task2026.03 | 81.5 | |
| One-Step VLATraining=Multi-task2026.03 | 64 | |
| Diffusion PolicyTraining=Single-task2026.03 | 46 |