You can deploy this model and start earning money today!

haohw/
behavior-uncloning-gd-step20

Name: behavior-uncloning-gd-step20
Availability: InStock

Views

No views yet

Behavior Uncloning — GD (step 20)

VLA unlearning checkpoint: pi0.5 model with GD unlearning applied.

Metric	Value
Method	GD
Training Steps	20
Forget Task	"turn on the stove" (LIBERO-Goal T6)
Forget SR	0% (baseline: 100%)
Retain SR	95.6% (baseline: 97.8%)
HM	0.98

bash

# Serve with openpi
uv run scripts/serve_policy.py --env LIBERO policy:checkpoint \
    --policy.config pi05_libero --policy.dir <path_to_checkpoint>

Gradient Difference: L = -L_forget + L_retain. Adds retain regularization to prevent catastrophic forgetting.