Skip to content

Add pi05_droid_jointpos config serving absolute joint actions #7

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
vertix merged 1 commit into from
Jul 19, 2026
Merged

Add pi05_droid_jointpos config serving absolute joint actions #7

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions src/openpi/training/config.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -711,6 +711,23 @@ def __post_init__(self) -> None:
),
),
),
# DROID in absolute joint-position action space (the RoboLab benchmark policies). The model predicts
# joint deltas; AbsoluteActions adds the observed joint position back, so the served chunk is absolute
# joint positions. Checkpoint: gs://openpi-assets-simeval/pi05_droid_jointpos.
TrainConfig(
name="pi05_droid_jointpos",
model=pi0_config.Pi0Config(action_horizon=15, pi05=True),
data=SimpleDataConfig(
assets=AssetsConfig(asset_id="droid"),
data_transforms=lambda model: _transforms.Group(
inputs=[droid_policy.DroidInputs(model_type=ModelType.PI05)],
outputs=[_transforms.AbsoluteActions(_transforms.make_bool_mask(7, -1)), droid_policy.DroidOutputs()],
),
base_config=DataConfig(
prompt_from_task=True,
),
),
),
#
# Fine-tuning Libero configs.
#
Expand Down
Loading