Open Source Python Reinforcement Learning Frameworks
Sort By:
Python Reinforcement Learning Frameworks
View 29 business solutionsBrowse free open source Python Reinforcement Learning Frameworks and projects below. Use the toggles on the left to filter open source Python Reinforcement Learning Frameworks by OS, license, language, programming language, and project status.
-
Gemini 3 and 200+ AI Models on One PlatformBuild, govern, and optimize agents and models with Gemini Enterprise Agent Platform.
-
MongoDB Atlas runs apps anywhereMongoDB Atlas gives you the freedom to build and run modern applications anywhere—across AWS, Azure, and Google Cloud. With global availability in over 115 regions, Atlas lets you deploy close to your users, meet compliance needs, and scale with confidence across any geography.
-
1
DeepSeek R1
Open-source, high-performance AI model with advanced reasoning
DeepSeek-R1 is an open-source large language model developed by DeepSeek, designed to excel in complex reasoning tasks across domains such as mathematics, coding, and language. DeepSeek R1 offers unrestricted access for both commercial and academic use. The model employs a Mixture of Experts (MoE) architecture, comprising 671 billion total parameters with 37 billion active parameters per token, and supports a context length of up to 128,000 tokens. DeepSeek-R1's training regimen uniquely integrates large-scale reinforcement learning (RL) without relying on supervised fine-tuning, enabling the model to develop advanced reasoning capabilities. This approach has resulted in performance comparable to leading models like OpenAI's o1, while maintaining cost-efficiency. To further support the research community, DeepSeek has released distilled versions of the model based on architectures such as LLaMA and Qwen. -
2
DeepSeek-V3
Powerful AI language model (MoE) optimized for efficiency/performance
DeepSeek-V3 is a robust Mixture-of-Experts (MoE) language model developed by DeepSeek, featuring a total of 671 billion parameters, with 37 billion activated per token. It employs Multi-head Latent Attention (MLA) and the DeepSeekMoE architecture to enhance computational efficiency. The model introduces an auxiliary-loss-free load balancing strategy and a multi-token prediction training objective to boost performance. Trained on 14.8 trillion diverse, high-quality tokens, DeepSeek-V3 underwent supervised fine-tuning and reinforcement learning to fully realize its capabilities. Evaluations indicate that it outperforms other open-source models and rivals leading closed-source models, achieving this with a training duration of 55 days on 2,048 Nvidia H800 GPUs, costing approximately $5.58 million. -
3
Agent S
Agent S: an open agentic framework that uses computers like a human
Agent S is an open-source agentic framework designed to enable autonomous computer use through an Agent-Computer Interface (ACI). Built to operate graphical user interfaces like a human, it allows AI agents to perceive screens, reason about tasks, and execute actions across macOS, Windows, and Linux systems. The latest version, Agent S3, surpasses human-level performance on the OSWorld benchmark, demonstrating state-of-the-art results in complex multi-step computer tasks. Agent S combines powerful foundation models (such as GPT-5) with grounding models like UI-TARS to translate visual inputs into precise executable actions. It supports flexible deployment via CLI, SDK, or cloud, and integrates with multiple model providers including OpenAI, Anthropic, Gemini, Azure, and Hugging Face endpoints. With optional local code execution, reflection mechanisms, and compositional planning, Agent S provides a scalable and research-driven framework for building advanced computer-use agents. -
4
H2O LLM Studio
Framework and no-code GUI for fine-tuning LLMs
Welcome to H2O LLM Studio, a framework and no-code GUI designed for fine-tuning state-of-the-art large language models (LLMs). You can also use H2O LLM Studio with the command line interface (CLI) and specify the configuration file that contains all the experiment parameters. To finetune using H2O LLM Studio with CLI, activate the pipenv environment by running make shell. With H2O LLM Studio, training your large language model is easy and intuitive. First, upload your dataset and then start training your model. Start by creating an experiment. You can then monitor and manage your experiment, compare experiments, or push the model to Hugging Face to share it with the community. -
Custom VMs From 1 to 96 vCPUs With 99.95% UptimeLive migration and automatic failover keep workloads online through maintenance. One free e2-micro VM every month.
-
5
Machine Learning PyTorch Scikit-Learn
Code Repository for Machine Learning with PyTorch and Scikit-Learn
Initially, this project started as the 4th edition of Python Machine Learning. However, after putting so much passion and hard work into the changes and new topics, we thought it deserved a new title. So, what’s new? There are many contents and additions, including the switch from TensorFlow to PyTorch, new chapters on graph neural networks and transformers, a new section on gradient boosting, and many more that I will detail in a separate blog post. For those who are interested in knowing what this book covers in general, I’d describe it as a comprehensive resource on the fundamental concepts of machine learning and deep learning. The first half of the book introduces readers to machine learning using scikit-learn, the defacto approach for working with tabular datasets. Then, the second half of this book focuses on deep learning, including applications to natural language processing and computer vision. -
6
Project Malmo
A platform for Artificial Intelligence experimentation on Minecraft
How can we develop artificial intelligence that learns to make sense of complex environments? That learns from others, including humans, how to interact with the world? That learns transferable skills throughout its existence, and applies them to solve new, challenging problems? Project Malmo sets out to address these core research challenges, addressing them by integrating (deep) reinforcement learning, cognitive science, and many ideas from artificial intelligence. The Malmo platform is a sophisticated AI experimentation platform built on top of Minecraft, and designed to support fundamental research in artificial intelligence. The Project Malmo platform consists of a mod for the Java version, and code that helps artificial intelligence agents sense and act within the Minecraft environment. The two components can run on Windows, Linux, or Mac OS, and researchers can program their agents in any programming language they’re comfortable with. -
7
CCZero (中国象棋Zero)
Implement AlphaZero/AlphaGo Zero methods on Chinese chess
ChineseChess-AlphaZero is a project that implements the AlphaZero algorithm for the game of Chinese Chess (Xiangqi). It adapts DeepMind’s AlphaZero method—combining neural networks and Monte Carlo Tree Search (MCTS)—to learn and play Chinese Chess without prior human data. The system includes self-play, training, and evaluation pipelines tailored to Xiangqi's unique game mechanics. -
8
Gymnasium
An API standard for single-agent reinforcement learning environments
Gymnasium is a fork of OpenAI Gym, maintained by the Farama Foundation, that provides a standardized API for reinforcement learning environments. It improves upon Gym with better support, maintenance, and additional features while maintaining backward compatibility. -
9
OSWorld
Benchmarking Multimodal Agents for Open-Ended Tasks
OSWorld is an open-source synthetic world environment designed for embodied AI research and multi-agent learning. It provides a richly simulated 3D world where multiple agents can interact, perform tasks, and learn complex behaviors. OSWorld emphasizes multi-modal interaction, enabling agents to process visual, auditory, and symbolic data for grounded learning in a simulated world. -
Stop Cyber Threats with VM-Series Next-Gen Firewall on AzureGain integrated visibility across all traffic in a single pass. Deploy Palo Alto Networks VM-Series to determine application identity and content while automating security policy updates via rich APIs.
-
10
Ray
A unified framework for scalable computing
Modern workloads like deep learning and hyperparameter tuning are compute-intensive and require distributed or parallel execution. Ray makes it effortless to parallelize single machine code — go from a single CPU to multi-core, multi-GPU or multi-node with minimal code changes. Accelerate your PyTorch and Tensorflow workload with a more resource-efficient and flexible distributed execution framework powered by Ray. Accelerate your hyperparameter search workloads with Ray Tune. Find the best model and reduce training costs by using the latest optimization algorithms. Deploy your machine learning models at scale with Ray Serve, a Python-first and framework agnostic model serving framework. Scale reinforcement learning (RL) with RLlib, a framework-agnostic RL library that ships with 30+ cutting-edge RL algorithms including A3C, DQN, and PPO. Easily build out scalable, distributed systems in Python with simple and composable primitives in Ray Core. -
11
TextWorld
TextWorld is a sandbox learning environment for the training
TextWorld is a learning environment designed to train reinforcement learning agents to play text-based games, where actions and observations are entirely in natural language. Developed by Microsoft Research, TextWorld focuses on language understanding, planning, and interaction in complex, narrative-driven environments. It generates games procedurally, enabling scalable testing of agents’ natural language processing and decision-making abilities. -
12
AnyTrading
The most simple, flexible, and comprehensive OpenAI Gym trading
gym-anytrading is an OpenAI Gym-compatible environment designed for developing and testing reinforcement learning algorithms on trading strategies. It simulates trading environments for financial markets, including stocks and forex. -
13
CORL
High-quality single-file implementations of SOTA Offline
CORL (Collection of Reinforcement Learning Environments for Control Tasks) is a modular and extensible set of high-quality reinforcement learning environments focused on continuous control and robotics. It aims to offer standardized environments suitable for benchmarking state-of-the-art RL algorithms in control tasks, including physics-based simulations and custom-designed scenarios. -
14
Catalyst
Accelerated deep learning R&D
Catalyst is a PyTorch framework for accelerated Deep Learning research and development. It allows you to write compact but full-featured Deep Learning pipelines with just a few lines of code. With Catalyst you get a full set of features including a training loop with metrics, model checkpointing and more, all without the boilerplate. Catalyst is focused on reproducibility, rapid experimentation, and codebase reuse so you can break the cycle of writing another regular train loop and make something totally new. Catalyst is compatible with Python 3.6+. PyTorch 1.1+, and has been tested on Ubuntu 16.04/18.04/20.04, macOS 10.15, Windows 10 and Windows Subsystem for Linux. It's part of the PyTorch Ecosystem, as well as the Catalyst Ecosystem which includes Alchemy (experiments logging & visualization) and Reaction (convenient deep learning models serving). -
15
Hands-on Unsupervised Learning
Code for Hands-on Unsupervised Learning Using Python (O'Reilly Media)
This repo contains the code for the O'Reilly Media, Inc. book "Hands-on Unsupervised Learning Using Python: How to Build Applied Machine Learning Solutions from Unlabeled Data" by Ankur A. Patel. Many industry experts consider unsupervised learning the next frontier in artificial intelligence, one that may hold the key to the holy grail in AI research, the so-called general artificial intelligence. Since the majority of the world's data is unlabeled, conventional supervised learning cannot be applied; this is where unsupervised learning comes in. Unsupervised learning can be applied to unlabeled datasets to discover meaningful patterns buried deep in the data, patterns that may be near impossible for humans to uncover. Author Ankur Patel provides practical knowledge on how to apply unsupervised learning using two simple, production-ready Python frameworks - scikit-learn and TensorFlow. With the hands-on examples and code provided, you will identify difficult-to-find patterns in data. -
16
Mctx
Monte Carlo tree search in JAX
mctx is a Monte Carlo Tree Search (MCTS) library developed by Google DeepMind for reinforcement learning research. It enables efficient and flexible implementation of MCTS algorithms, including those used in AlphaZero and MuZero. -
17
WikiSQL
A large annotated semantic parsing corpus for developing NL interfaces
A large crowd-sourced dataset for developing natural language interfaces for relational databases. WikiSQL is the dataset released along with our work Seq2SQL: Generating Structured Queries from Natural Language using Reinforcement Learning. Regarding tokenization and Stanza, when WikiSQL was written 3-years ago, it relied on Stanza, a CoreNLP python wrapper that has since been deprecated. If you'd still like to use the tokenizer, please use the docker image. We do not anticipate switching to the current Stanza as changes to the tokenizer would render the previous results not reproducible. -
18
dm_control
DeepMind's software stack for physics-based simulation
DeepMind's software stack for physics-based simulation and Reinforcement Learning environments, using MuJoCo. DeepMind's software stack for physics-based simulation and Reinforcement Learning environments, using MuJoCo physics. The MuJoCo Python bindings support three different OpenGL rendering backends: EGL (headless, hardware-accelerated), GLFW (windowed, hardware-accelerated), and OSMesa (purely software-based). At least one of these three backends must be available in order render through dm_control. Hardware rendering with a windowing system is supported via GLFW and GLEW. On Linux these can be installed using your distribution's package manager. "Headless" hardware rendering (i.e. without a windowing system such as X11) requires EXT_platform_device support in the EGL driver. While dm_control has been largely updated to use the pybind11-based bindings provided via the mujoco package, at this time it still relies on some legacy components that are automatically generated. -
19
gym-pybullet-drones
PyBullet Gymnasium environments for multi-agent reinforcement
Gym-PyBullet-Drones is an open-source Gym-compatible environment for training and evaluating reinforcement learning agents on drone control and swarm robotics tasks. It leverages the PyBullet physics engine to simulate quadrotors and provides a platform for studying control, navigation, and coordination of single and multiple drones in 3D space. -
20
Acme
A library of reinforcement learning components and agents
Acme is a framework from DeepMind for building scalable and reproducible reinforcement learning agents. It emphasizes modular components, distributed training, and ease of experimentation. -
21
AgentUniverse
agentUniverse is a LLM multi-agent framework
AgentUniverse is a multi-agent AI framework that enables coordination between multiple intelligent agents for complex task execution and automation. -
22
Alibi Explain
Algorithms for explaining machine learning models
Alibi is a Python library aimed at machine learning model inspection and interpretation. The focus of the library is to provide high-quality implementations of black-box, white-box, local and global explanation methods for classification and regression models. -
23
AndroidEnv
RL research on Android devices
android_env is a reinforcement learning (RL) environment developed by Google DeepMind that enables agents to interact with Android applications directly as a learning environment. It provides a standardized API for training agents to perform tasks on Android apps, supporting tasks ranging from games to productivity apps, making it suitable for research in real-world RL settings. -
24
Astrape
Optical-packet node transceiver frequency allocation
In an optical network scenario which consists of multiple nodes (whiteboxes) at its edges and ROADMs in-between, the coherent transceiver average laser configuration time is improved. The process is evaluated according to a testbed setup. This is facilitated in the appropriate lab equipment (or via simulation when required). For that purpose, a software agent (Netconf server) residing at the whiteboxes, is developed receiving input from the Software-Defined Networking (SDN) packet controller (PacketCTL - a Netconf client). Then, configuration of the local transceiver laser frequencies of the controlled pluggable devices takes place, for facilitating the connectivity in-between the ROADM network. Also, the agent records and reports back telemetry data (feedback) which is used by the PacketCTL's resource-allocating mechanism to improve efficiency within the network topology. -
25
Atropos
Language Model Reinforcement Learning Environments frameworks
Atropos is a comprehensive open-source framework for reinforcement learning (RL) environments tailored specifically to work with large language models (LLMs). Designed as a scalable ecosystem of environment microservices, Atropos allows researchers and developers to collect, evaluate, and manage trajectories (sequences of actions and outcomes) generated by LLMs across a variety of tasks—from static dataset benchmarks to dynamic interactive games and real-world scenario environments. It provides foundational tooling for asynchronous RL loops where environment services communicate with trainers and inference engines, enabling complex workflow orchestration in distributed and parallel setups. This framework facilitates experimentation with RLHF (Reinforcement Learning from Human Feedback), RLAIF, or multi-turn training approaches by abstracting environment logic, scoring, and logging into reusable components.
