bryce301best
/

fold_towel

imitation-learning

relative-actions

Model card Files Files and versions

PI0.5 Fold Towel Relative-Action Checkpoints

This repository stores PI0.5 fine-tuning checkpoints for chenglongy/fold_towel using LeRobot relative actions.

Layout

checkpoints/002000/pretrained_model
checkpoints/004000/pretrained_model
checkpoints/006000/pretrained_model
checkpoints/008000/pretrained_model
checkpoints/010000/pretrained_model
checkpoints/012000/pretrained_model
eval/

The training_state/ optimizer and scheduler state is intentionally omitted. These files are for inference, evaluation, and checkpoint comparison.

Notes

Policy type: pi05
Relative actions: enabled
Relative exclude joints: gripper

Downloads last month: -; Downloads are not tracked for this model. How to track

Video Preview

loading