MPTransport

class torchrl.modules.inference_server.MPTransport(ctx: BaseContext | None = None)

Cross-process transport using multiprocessing queues.

Response routing uses per-actor queues (one per client() call) so that no mp.Queue object is ever serialised through another queue. Clients must be created with client() before spawning child processes.

Parameters:

ctx – a multiprocessing context (e.g. mp.get_context("spawn")). Defaults to mp.get_context("spawn").

Example

>>> import multiprocessing as mp
>>> transport = MPTransport()
>>> client = transport.client()           # creates response queue
>>> p = mp.Process(target=actor_fn, args=(client,))
>>> p.start()                             # queue inherited

client() → _QueueInferenceClient

Create an actor-side client with a dedicated response queue.

Must be called in the parent process before spawning children.

Returns:

A _QueueInferenceClient that can be passed to a child process as an argument to multiprocessing.Process.