Visual Grounding on Open-World (test)
73Success RateOneTwoVLA-VL
Evaluation results
| Method | Links | |
|---|---|---|
| OneTwoVLA-VLTraining Data=Robot data + 16,000 synthetic vision-language samples2025.05 | 73 | |
| OneTwoVLATraining Data=Exclusively robot data2025.05 | 8 | |
| π0Training Data=Exclusively robot data2025.05 | 3 |