Handover Apple on RoboTwin
86Success Rateπ0.5 + PACT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| π0.5 + PACTPost-training=PACT, Fine-tuning Strategy=LoRA2026.06 | 86 | 61 | |
| RDT-1B + PACTPost-training=PACT, Fine-tuning Strategy=LoRA2026.06 | 83 | 59 | |
| DP + PACTPost-training=PACT, Fine-tuning Strategy=Full-parameter2026.06 | 81 | 82 | |
| RDT-1BPost-training=None2026.06 | 76 | 52 | |
| π0.5Post-training=None2026.06 | 71 | 55 | |
| DP3 + PACTPost-training=PACT, Fine-tuning Strategy=Full-parameter2026.06 | 70 | 59 | |
| DPPost-training=None2026.06 | 63 | 64 | |
| DP3Post-training=None2026.06 | 54 | 46 |