reinforcement learning: 41 curated resources
reinforcement learning
41 resources · page 1 of 2
AlphaChip (circuit_training)
Google Research circuit_training GitHub repo: AlphaChip framework for chip floorplanning with reinforcement learning.
AlphaDev
DeepMind AlphaDev GitHub repo: pseudocode and algorithms from the faster sorting algorithms paper.
google-deepmind/alphatensor
DeepMind AlphaTensor GitHub repo: code for discovering faster matrix multiplication algorithms with RL.
Awex
by inclusionAI
inclusionAI Awex GitHub repo: weight synchronisation between RL training and inference engines.
DeepSeek-R1
by DeepSeek
DeepSeek-R1 GitHub repo: model summary, evaluation and usage for DeepSeek-R1 and R1-Zero reasoning models.
FinRL
FinRL GitHub repo (AI4Finance): financial reinforcement learning framework for trading.
Flow
by flow-project
Flow (UC Berkeley): framework combining traffic simulation and deep RL for mixed-autonomy traffic control.
Gymnasium
Gymnasium documentation: standard API and reference environments for reinforcement learning (formerly OpenAI Gym).
Mava
InstaDeep Mava GitHub repo: codebase for experiments in multi-agent reinforcement learning.
MS-SWIFT
by ModelScope
ModelScope ms-swift GitHub repo: framework for PEFT and full fine-tuning, DPO and GRPO of 600+ LLMs.
OpenAI Spinning Up
by OpenAI
OpenAI Spinning Up: educational guide to deep reinforcement learning with algorithm explanations and code.
OpenRLHF
by OpenRLHF
OpenRLHF GitHub repo: Ray and vLLM based framework for RLHF and agentic RL training.
OpenSpiel
DeepMind OpenSpiel GitHub repo: environments and algorithms for reinforcement learning and search in games.
slime
by THUKEG
THUDM slime GitHub repo: Megatron and SGLang based framework for RL post-training at scale, used for GLM.
THUDM (Tsinghua KEG) GitHub
by THUKEG
THUDM GitHub organisation (Tsinghua KEG and Zhipu): code for ChatGLM, GLM-4, CogVLM, CogVideoX and AgentBench.
verl
by verl-project
verl GitHub repo (originally ByteDance): RL post-training framework used for RLVR.
verl / HybridFlow
by verl-project
verl GitHub repo: HybridFlow framework for distributed RL post-training of LLMs.
HF Deep RL Course
by Hugging Face
Hugging Face Deep Reinforcement Learning Course: free course on RL theory and practice with hands-on units.
Unsloth RL / GRPO guide
by Unsloth AI
Unsloth reinforcement learning guide: training reasoning models with GRPO, with linked tutorials and advanced RL docs.
AlphaGo
Google DeepMind AlphaGo page covering the Go-playing system, its matches and the research papers behind it.
AlphaStar: Mastering the real-time strategy game StarCraft II
Google DeepMind post on AlphaStar, the reinforcement-learning system that reached Grandmaster level in StarCraft II.
Easy-RL / 蘑菇书
by Datawhale
Datawhale Easy-RL GitHub repo: Chinese reinforcement-learning tutorial (Mushroom Book) with code.
YSDA Practical RL
by Yandex School of Data Analysis
YSDA Practical_RL GitHub repo: open reinforcement-learning course with notebooks and assignments from HSE and YSDA.
Agent57
by Adrià Puigdomènech Badia et al. (Google)
arXiv paper (2020) introducing Agent57, the first agent to exceed the human baseline on all 57 Atari games.
Faster sorting algorithms discovered using deep reinforcement learning (AlphaDev)
by Daniel J. Mankowitz et al. (Google DeepMind)
Nature paper (2023) on AlphaDev discovering faster sorting algorithms with deep RL.
Mastering the game of Go with deep neural networks and tree search (AlphaGo)
by David Silver et al. (DeepMind)
Nature paper (2016) on AlphaGo, mastering Go with deep neural networks and tree search.
Grandmaster level in StarCraft II using multi-agent reinforcement learning (AlphaStar)
by Oriol Vinyals et al. (DeepMind)
Nature paper (2019) on AlphaStar reaching grandmaster level in StarCraft II with multi-agent RL.
Discovering faster matrix multiplication algorithms with reinforcement learning (AlphaTensor)
by Alhussein Fawzi et al. (Google DeepMind)
Nature paper (2022) on AlphaTensor discovering faster matrix multiplication algorithms with RL.
AlphaZero
by David Silver et al. (DeepMind)
arXiv paper (2017) introducing AlphaZero, mastering chess and shogi through self-play reinforcement learning.
DQN (Atari)
by Volodymyr Mnih et al. (Google DeepMind)
arXiv paper (2013) "Playing Atari with Deep Reinforcement Learning", introducing DQN, deep Q-learning from pixels on Atari games.
DreamerV3
by Danijar Hafner et al. (Google DeepMind)
arXiv paper (2023) introducing DreamerV3, a world-model RL agent that masters diverse domains with fixed settings.
IMPALA
by Lasse Espeholt et al. (Google)
arXiv paper (2018) introducing IMPALA, a scalable distributed deep RL architecture.