Rl Python Packages

torchrl

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

sb3-contrib

Contrib package for Stable-Baselines3 - Experimental reinforcement learning (RL) code

tensorlake

Tensorlake is a serverless runtime for sandboxes and deploying background agentic applications

tianshou

An elegant PyTorch deep reinforcement learning library.

judgeval

The Continuous-Improvement Stack for Agents. Our environment data and evals power agent improvement and monitoring.

gem-llm

A Gym for Agentic LLMs

neptune

📘 The experiment tracker for foundation model training

dopamine-rl

Dopamine is a research framework for fast prototyping of reinforcement learning algorithms.

neptune-client

📘 The experiment tracker for foundation model training

hud-python

RL environments + evals for AI agents. Define once, train anything.

torchrl-nightly

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

rliable

[NeurIPS'21 Outstanding Paper] Library for reliable evaluation on RL and ML benchmarks, even with only a handful of seeds.

torchstudio

Deep Learning Experiment

flashbax

⚡ Flashbax: Accelerated Replay Buffers in JAX

rl-zoo3

A training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.

navix

Accelerated minigrid environments with JAX

multi-agent-rlenv

Strongly typed reinforcement learning environment framework

amp-rsl-rl

🔁 AMP-RSL-RL: Adversarial Motion Priors for robotic RL (PPO + motion imitation)

awex

A high-performance RL training-inference weight synchronization framework, designed to enable second-level parameter updates from training to inference in RL workflows

dsse

The Drone Swarm Search project provides an environment for SAR missions built on PettingZoo, where agents, represented by drones, are tasked with locating targets identified as shipwrecked individuals.

trianglengin

High-performance C++/Python engine for a triangle puzzle game.

moonfish

~2000 Elo Python Chess Engine that implements: Negamax, PeSTO’s Evaluation, Null Move, Quiescence Search, Lazy SMP.

rlox

Rust-accelerated reinforcement learning — 22 algorithms, 2-10x faster than SB3. The Polars of RL.

areno

An easy-to-use, fast toolkit to scale up RL post-training on a single node.