openvinotoolkit/physicalai/pai-load-policy

Loading Exported Policies

Runtime loads Studio export directories (or Hub snapshots that mirror them) through InferenceModel. It reads manifest.json, auto-detects registered ONNX/OpenVINO adapters from model files, and supports Hub revisions. See the inference API and manifest schema.

Workflow

1. Identify the artifact: local export directory or Hub repo_id, expected backend, and whether the user needs select_action vs predict_action_chunk.

• Done when: load path and API entry point are chosen before editing code.

1. Load with auto-detection first (local):

``python from physicalai.inference import InferenceModel model = InferenceModel("./exports/act_policy") ``

• Done when: model constructs without explicit backend= when artifacts match a registered extension.

1. Hub load when the package is published:

``python model = InferenceModel.from_pretrained("OpenVINO/act-fp16-ov", revision="<commit-sha>") ``

• Done when: revision is pinned for reproducibility when security or CI matters.

1. Explicit backend only when auto-detection is ambiguous:

``python model = InferenceModel("./exports/act_policy", backend="openvino", device="CPU") ``

1. Validate structure against the manifest schema and backend notes (references/onnx.md, references/openvino.md).

• Done when: manifest, model file, and processor artifacts resolve under the export directory.

1. Smoke inference without owning robot timing:

``python model.reset() action = model.select_action(observation) ``

• Done when: one forward pass succeeds on representative observation keys/shapes. For hardware loops, hand off to pai-run-policy.

Validation loop

Construct InferenceModel from the export and run one representative fake observation without hardware. If a backend cannot be loaded, check its installed optional dependency and the documented adapter support.

Required checks

• Export directory contains backend model file(s) and manifest.json consistent with docs/reference/manifest-schema.md.
• Metadata input/output/feature names align with preprocessors in the manifest.
• Optional backends fail with clear install guidance (onnxruntime, OpenVINO).
• Do not claim support for an adapter absent from the installed Runtime package.
• Hub loads must not log tokens; prefer pinned revision= for production docs.

References

• Manifest schema — canonical export/load contract.
• references/onnx.md, references/openvino.md — adapter constraints.