--- license: apache-2.0 base_model: lerobot/VLA-JEPA-Pretrain pipeline_tag: robotics tags: - LeRobot - so101 - vla-jepa - lora --- # vla_jepa_so101_block_stack_sim VLA-JEPA fine-tuned to stack one 30 mm block on another with an [SO-101](https://github.com/TheRobotStudio/SO-ARM100), in simulation. Trained by `train/vla_jepa_so101.yaml` in [vla-test](https://github.com/danwahl/vla-test) on [drwahl/so101_block_stack_sim](https://huggingface.co/datasets/drwahl/so101_block_stack_sim). LoRA (r=32, alpha 64) on the Qwen language backbone of `lerobot/VLA-JEPA-Pretrain`, with the action model and the video predictor dense, vision frozen: 20k steps at batch 4, the protocol [drwahl/pi05_so101_block_stack_sim](https://huggingface.co/drwahl/pi05_so101_block_stack_sim) was trained on. Normalization, learning rate and chunk length are the base checkpoint's own. `adapter/` holds the adapter and the processors, which carry the normalization and the camera renaming; lerobot loads it onto the base weights. ```bash hf download drwahl/vla_jepa_so101_block_stack_sim --local-dir vla_jepa_so101_block_stack_sim uv run python sim/scripts/eval.py vla_jepa_so101_block_stack_sim/adapter --no-smooth ``` The eval rolls the policy out on the 150 held-out layouts and scores it by the gates that screened the oracle. Executing its whole 7-step chunk, the setting that stacks most on one seed, it stacks 91.7 of 150 and lifts 105.7, averaged over three further seeds. A 5-step horizon stacks 59 on one seed against 93 for the whole chunk.