gs://openpi-assets/checkpoints/pi0_base) on the
LIBERO benchmark (physical-intelligence/libero dataset), using the
openpi pi0_libero training config.pi0_basephysical-intelligence/liberoparams/ (inference weights) and assets/ (norm stats) needed to run inference —
the optimizer/train_state/ used to resume training is not included.1from openpi.training import config as _config
2from openpi.policies import policy_config
3
4cfg = _config.get_config("pi0_libero")
5policy = policy_config.create_trained_policy(cfg, "<path to this checkpoint>")