Vector Wang

Currently not directly available code. But you can get started by following Huggingface LeRobot official VLA tutorials, realize it on single arm first and then it should be not hard...

Directly transferring is kind of difficult since you need to take care of the power and control exclusively for the larger arms, which is not directly compatible to LeRobot. I...

Previously I was using D415, but D435 or D435i should also works. I think the kit provides two kinds of head camera mounts, one should be able to mount Realsense...

Definitely. Right now the Maniskill simulation has all the teleop codes and sensory feedback. And we are actively working on the Maniskill hab environment for RL mobile manipulation and VLA....

Yes, I have just quickly written some codes that can record the simulation in the latest lerobot dataset v3 version. You can also change the controller or modify the recording...

头部再高的话移动过程中的晃动会过大，目前新设计的底座的头部相机更靠前了一些（见hardware/ongoing），遮挡情况会好一些。目前市面上绝大多数机器人应该也都无法避免这个问题。 想彻底解决还是推荐在训练时充分考虑手部相机输入，以及必要的话在桌面上增加额外的相机。

Thanks for your detailed suggestions! We will release a assembly tutorial video soon to address some of these issues. After that I will further polish the text tutorial for the...

理论上这个代码应该是可以自动下载yolo模型的，可以先试试看另一个example（_follow）看看是不是有同样的问题。 下面是AI的回答： 我看到了你的错误是 PytorchStreamReader failed reading zip archive: failed finding central directory，通常表示“权重文件不是一个有效的 PyTorch zip/JIT 文件或已损坏”。绝大多数时候是权重下载不完整/格式不匹配导致的。按下面步骤逐一排查，基本能定位并解决。 ### 快速定位（先验证环境没问题） - 用一个已知可用的权重试加载（你目录下已有 `yolo11n-seg.pt`/`yolo11l-seg.pt`）： ```python from ultralytics import YOLO YOLO("/home/vector/Research/XLeRobot/yolo11n-seg.pt") # 或 yolo11l-seg.pt ```...

目前已实现的demo里面不涉及导航，在RL demo中机器人直接在局部区域进行RL训练并直接通过处理过的视觉信息实现底盘动作。未来会基于目前前沿的VLN，深度SLAM，双目slam等方法在不涉及雷达的情况下通过深度或双目相机对环境空间进行全局建模

I just downloaded the file again and it can open normally. Maybe your download broke in the middle or something?