Run a Policy on a Robot

Preview: PolicyRuntime and the CLI are planned APIs. The examples below document the target design.

Python API

from physicalai.runtime import PolicyRuntime, SyncExecution
from physicalai.inference import InferenceModel
from physicalai.robot import SO101
from physicalai.capture import UVCCamera

runtime = PolicyRuntime(
    fps=30,
    robot=SO101(port="/dev/ttyACM0"),
    model=InferenceModel.load("./exports/act_policy"),
    cameras={
        "wrist": UVCCamera(device="/dev/video0", width=640, height=480),
    },
    execution=SyncExecution(mode="chunk"),
)

runtime.run(duration_s=60)

From Config

Write a runtime configuration file.

# runtime.yaml
runtime:
  class_path: physicalai.runtime.PolicyRuntime
  init_args:
    fps: 30
    robot:
      class_path: physicalai.robot.so101.SO101
      init_args:
        port: /dev/ttyACM0
    model:
      class_path: physicalai.inference.InferenceModel
      init_args:
        export_dir: ./exports/act_policy
    cameras:
      wrist:
        class_path: physicalai.capture.UVCCamera
        init_args:
          device: /dev/video0
          width: 640
          height: 480
    execution:
      class_path: physicalai.runtime.SyncExecution
      init_args:
        mode: chunk

Load and run from Python.

from physicalai.runtime import PolicyRuntime

runtime = PolicyRuntime.from_config("runtime.yaml")
runtime.run(duration_s=60)

Or run from the CLI.

physicalai run --config runtime.yaml --duration-s 60

Component Responsibilities

Object	Owns
InferenceModel	policy inference
PolicyRuntime	robot loop and timing
Execution	where inference runs
Robot	hardware IO
Camera	image capture