pi0_busybox_push_green_button

Full-component fine-tune of π0 for busybox_push_green_button on SO101 data.

Policy π0 (pi05=false, full-component)
Init checkpoint pi0_base (gs://openpi-assets/checkpoints/pi0_base)
Dataset villekuosmanen/busybox_push_green_button
Task busybox_push_green_button
Prompt push the green button
Action dim 6 (single-arm)
Cameras top → base_0_rgb, wrist → left_wrist_0_rgb, front → base_1_rgb
Training GMAN 4× H100 80GB data-parallel, full-component, 30k steps, global batch 32, action horizon 30, 720x1280 cameras
W&B project busybox_push_green_button_pi0
W&B run q1f9flg5

Checkpoints

main is the latest finalized checkpoint (params/ + assets/ at the repo root). The publisher overwrites the root at each 5k save. Training target is step 30,000.

Field Value
Latest published main (params/)
Config pi0_busybox_push_green_button

Usage

from huggingface_hub import snapshot_download
from openpi.policies import policy_config
from openpi.training.config import get_config

ckpt = snapshot_download("pravsels/pi0_busybox_push_green_button")
policy = policy_config.create_trained_policy(
    get_config("pi0_busybox_push_green_button"),
    ckpt,
)
uv run scripts/serve_policy.py \
  --policy.config=pi0_busybox_push_green_button \
  --policy.dir=/path/to/pravsels/pi0_busybox_push_green_button
Downloads last month

-

Downloads are not tracked for this model. How to track
Video Preview
loading

Dataset used to train pravsels/pi0_busybox_push_green_button