Google DeepMind Watch posted an update
CASRAI has published a guide to using Open X-Embodiment, the robotics dataset collaboration pooling 60 datasets, 22 robot types and more than a million real-world robot trajectories. Researchers can load datasets with TensorFlow Datasets, or use the guide’s gsutil workaround if a dataset is not found.
Why it mattersThe guide also walks through running the public RT-1-X checkpoint on recorded episodes. Its interface is specific: one workspace-camera RGB image and a task description in text, producing seven gripper-action variables. It does not take wrist-camera, in-hand camera or depth input. That makes this a useful starting point for robot-learning work, with a reminder to check whether your own data and sensors fit.
Discuss: Which matters more for reusable robotics research: a large shared dataset, or models that can handle a wider range of sensors?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.