Gespeichert in:
| Hauptverfasser: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2509.00576 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915472303194112 |
|---|---|
| author | Jiang, Tao Yuan, Tianyuan Liu, Yicheng Lu, Chenhao Cui, Jianning Liu, Xiao Cheng, Shuiqi Gao, Jiyang Xu, Huazhe Zhao, Hang |
| author_facet | Jiang, Tao Yuan, Tianyuan Liu, Yicheng Lu, Chenhao Cui, Jianning Liu, Xiao Cheng, Shuiqi Gao, Jiyang Xu, Huazhe Zhao, Hang |
| contents | We present Galaxea Open-World Dataset, a large-scale, diverse collection of robot behaviors recorded in authentic human living and working environments. All demonstrations are gathered using a consistent robotic embodiment, paired with precise subtask-level language annotations to facilitate both training and evaluation. Building on this dataset, we introduce G0, a dual-system framework that couples a Vision-Language Model (VLM) for multimodal planning with a Vision-Language-Action (VLA) model for fine-grained execution. G0 is trained using a three-stage curriculum: cross-embodiment pre-training, single-embodiment pre-training, and task-specific post-training. A comprehensive benchmark spanning tabletop manipulation, few-shot learning, and long-horizon mobile manipulation, demonstrates the effectiveness of our approach. In particular, we find that the single-embodiment pre-training stage, together with the Galaxea Open-World Dataset, plays a critical role in achieving strong performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_00576 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Galaxea Open-World Dataset and G0 Dual-System VLA Model Jiang, Tao Yuan, Tianyuan Liu, Yicheng Lu, Chenhao Cui, Jianning Liu, Xiao Cheng, Shuiqi Gao, Jiyang Xu, Huazhe Zhao, Hang Robotics Computer Vision and Pattern Recognition We present Galaxea Open-World Dataset, a large-scale, diverse collection of robot behaviors recorded in authentic human living and working environments. All demonstrations are gathered using a consistent robotic embodiment, paired with precise subtask-level language annotations to facilitate both training and evaluation. Building on this dataset, we introduce G0, a dual-system framework that couples a Vision-Language Model (VLM) for multimodal planning with a Vision-Language-Action (VLA) model for fine-grained execution. G0 is trained using a three-stage curriculum: cross-embodiment pre-training, single-embodiment pre-training, and task-specific post-training. A comprehensive benchmark spanning tabletop manipulation, few-shot learning, and long-horizon mobile manipulation, demonstrates the effectiveness of our approach. In particular, we find that the single-embodiment pre-training stage, together with the Galaxea Open-World Dataset, plays a critical role in achieving strong performance. |
| title | Galaxea Open-World Dataset and G0 Dual-System VLA Model |
| topic | Robotics Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.00576 |