sourccey-shirt-fold-010

LeRobot xVLA policy trained for shirt folding using the Combination/sourccey-shirt-fold-c-010 dataset.

Checkpoint

This repository contains training checkpoint 500000:

• pretrained_model/ contains the policy configuration, weights, and processors used for inference.
• training_state/ contains optimizer, scheduler, and random-number-generator state for resuming training.

The policy consumes a 16-dimensional state and four 240 x 320 camera views (front left, front right, left wrist, and right wrist) and predicts a 16-dimensional action.

Limitations

This checkpoint should be validated on the target robot and workspace before deployment. Its behavior is limited by the coverage and quality of its training dataset.