openpi π₀.₅ policy checkpoint.
-
Train config: pi05_Franka_EXP8_10
-
Experiment: EXP8_10_a100
-
Checkpoints in this repo:
-
step_2999/
Each folder is one training step and is self-contained.
1from huggingface_hub import snapshot_download
2from openpi.policies import policy_config
3from openpi.training import config as _config
4
5ckpt = snapshot_download("saifahmad123/EXP8_10_a100", allow_patterns="step_2999/*") + "/step_2999"
6train_config = _config.get_config("pi05_Franka_EXP8_10")
7policy = policy_config.create_trained_policy(train_config, ckpt)
8
9action_chunk = policy.infer(observation)["actions"]
