agents.clients.lerobot

Module Contents

Classes

LeRobotClient

A GRPC client for interaction with robot action policies served on LeRobot

API

class agents.clients.lerobot.LeRobotClient(model: Union[agents.models.LeRobotPolicy, Dict], host: str = '127.0.0.1', port: int = 8080, inference_timeout: int = 30, init_on_activation: bool = True, logging_level: str = 'info', **kwargs)

Bases: agents.clients.model_base.ModelClient

A GRPC client for interaction with robot action policies served on LeRobot

receive_actions()

Receive actions from the server

serialize() → Dict

Get client json

Return type:

Dict

property supports_tool_calls: bool

Check if the client supports tool calling (function calling). Defaults to False. Override in subclasses if supported.

Return type:

bool

property context_length: Optional[int]

The context length, in tokens, the model is served with. Defaults to None. Override in subclasses that can report it.

Return type:

Optional[int]

property supports_decisions: bool

Check if the client serves a decision model, which answers typed questions about a state. Defaults to False. Override in subclasses if supported.

Return type:

bool

check_connection() → None

initialize.

Return type:

None

initialize() → None

initialize.

Return type:

None

inference(inference_input: Dict[str, Any]) → Optional[MutableMapping]

inference.

Parameters:

inference_input (dict[str, Any])

Return type:

dict | None

deinitialize()

deinitialize.