rlmesh 0.1.0rc3

Python SDK for model-environment evaluation workflows

RLMesh

RLMesh is a Python SDK for model-environment evaluation. It serves Gymnasium-style environments, connects clients over local or remote transports, and adapts values for plain Python, NumPy, Torch, and JAX users.

Pre-1.0 (0.x): the stable API may change in a minor release, with a migration note, so pin a minor range for active projects.

Installation

Install from PyPI:

pip install rlmesh

Install optional adapters as needed:

pip install "rlmesh[numpy]"
pip install "rlmesh[gymnasium]"
pip install "rlmesh[torch]"
pip install "rlmesh[jax]"

Quickstart

Install RLMesh with Gymnasium support and the NumPy client adapter:

pip install "rlmesh[gymnasium,numpy]"

In one process, serve any Gymnasium-compatible environment:

import gymnasium as gym
from rlmesh import EnvServer

env = gym.make("CartPole-v1")
EnvServer(env, "127.0.0.1:5555").serve()

In another process, connect to it as a remote environment:

from rlmesh.numpy import RemoteEnv

env = RemoteEnv("127.0.0.1:5555")
observation, info = env.reset(seed=0)

terminated = truncated = False
while not (terminated or truncated):
    action = env.action_space.sample()
    observation, reward, terminated, truncated, info = env.step(action)

env.close()

Links

• Homepage: https://rlmesh.dev
• Documentation: https://docs.rlmesh.dev
• Repository: https://github.com/ArenaX-Labs/rlmesh
• Examples: https://github.com/ArenaX-Labs/rlmesh/tree/main/examples/python
• Issues: https://github.com/ArenaX-Labs/rlmesh/issues
• Default contact: [e-mail]

License

Licensed under either of Apache License, Version 2.0 or the MIT license, at your option. See LICENSE-APACHE and LICENSE-MIT.

Python wheels also include third-party notices in THIRD_PARTY_NOTICES.md.