| This is the model checkpoint of "RoboAgent: Chaining Basic Capabilities for Embodied Task Planning" (CVPR 2026). | |
| It is a [Qwen2.5-VL-3B](https://github.com/qifan-han/Qwen2.5-VL) finetuned on [ALFRED](https://github.com/askforalfred/alfred)'s training set. | |
| paper: https://arxiv.org/abs/2604.07774 | |
| code: https://huggingface.co/woyut/RoboAgent_CVPR26 |