pi05-pick3-baseline-lora-gbs4-pgb1-gpu4-r2-step12500

Second-round π0.5 LoRA fine-tuning for pick3, checkpoint 12,500. Trained locally on 4 RTX A6000 GPUs, global batch 4, per-GPU batch 1, without gradient accumulation. This checkpoint represents 50,000 sample exposures; the run targets 12,500 optimizer updates / 50,000 sample exposures. Checkpoints are saved every 1,000 updates and at the final 12,500. Steps 5,000, 10,000, and 12,500 have standalone model repositories. Other complete checkpoints are retained in the backup archive under task/method/stepN/ with the same loading files and full resume state. r2 keeps these artifacts separate from the original 6,250-step experiment.

Warmup: 250 updates to 5e-5; cosine decay to 5e-6 at step 12,500. EMA: 0.999^4 = 0.996005996001. AdamW, LoRA rank32, seed42. Full settings are in training_config.json.

No-history baseline: current base/wrist images, state and task instruction.

Only robot execution rows supervise actions. Demonstration images, features and original episode/frame coordinates remain available as history; execution start does not reset the visual history. Episode splits and normalization use the training split only. The recorded Cartesian end-effector/tool pose convention must match the collection controller; do not apply an extra tool/flange offset. Outputs have shape (20,8): absolute xyz, unit XYZW quaternion in the positive-qx chart, and gripper command in [0,1]. Numeric state/actions use STD normalization; state tokens use a separate bounded train-q01/q99 view.

Unknown action components remain NaN/false in the training dataset and are masked in both flow conditioning and loss. Entirely unsupervised components: []. In particular, Shuffle's gripper output has no command supervision and must not be interpreted as learned gripper control. Button Order has only limited verified closed-command labels.

This repository includes EMA serving weights, normalization/history assets, inference source, exact dependency versions, and full non-EMA/optimizer/RNG/sampler resume state. Intermediate publication is not a policy-quality evaluation. Offline evaluation, when included, does not establish real-robot success rate.

Load with from load_model import load, observe; policy = load() from this bundle. For history models call observe(policy, base_rgb, state) for every observed frame, including demonstrations, and reset between episodes. Status-D additionally requires explicit history_keyframe_index (an observed frame or None), current_subgoal, and causal transition_context_* inputs in policy.infer(...); its output includes transition_status. Use the exported Writer schedule and matching control frame.

Final checkpoint checks: offline evaluation completed; packaged-source inference and finite-output checks passed. See evaluation.json and packaged-inference-check.txt; no physical robot success-rate claim.