vla_real_pk_remove_sharp_full_step550
Edited pi0.5 VLA checkpoint for pass_knife task — pk_remove_sharp_full arm at step 550.
Deployment goal
Remove unsafe 'sharp' (no-rotation) handoff behavior.
Edit recipe
- steering_mode: hidden_v9_mc_softhybrid_precommit_gated
- target_subset: {0:left, 1:right}
- loss_formula: -logsumexp(z_target_subset) + logsumexp(z_others)
- gating: frame_index < 999 (full trajectory)
- ablation_arm: FULL (control)
Common hyperparameters
γ=0.1, β=1.0, lr=1e-5, batch=32, num-steps=400-600, save-interval=50-100, ViT frozen
Foundation VLA
pi05_real_pk_mixed/real_pk_mixed_v3 step 24999 — frozen mixed-mode foundation, edited only on the action-expert + LLM (ViT frozen).
Classifier used
/mnt/data3/classifiers/real_v3/pk_v5h_mc/best_v5h_mc.pt (v5h-mc 3-class softmax, val_acc 91-94%)
Why this checkpoint?
Top-2 by composite val_loss_pref + 0.1 * loss_redirect.
The composite score balances target-mode preservation (val_loss_pref low) and active editing pressure (loss_redirect strongly negative).
For all 4 main edits, the LATEST ckpts won by composite score — i.e., the editing benefits accumulate throughout training and don't overfit before the final step in this configuration.
Ablation companion
pk_remove_sharp_gradgate — eval both and compare to validate the gating choice.
Eval target
50-seed real-robot rollouts. Compare:
- Target rate: fraction of episodes where deployment goal is achieved
- Overall SR: model still completes the task successfully
- vs foundation VLA baseline (no editing) on the same seeds
Loading
1from openpi.training import config as _config
2import openpi.shared.array_typing as at
3from openpi.models.model import preprocess_observation
4# Load this repo's params/ subdirectory as the model checkpoint.
5# Use config: pi05_real_pk_mixed
