Python deep-reinforcement-learning

Open-source Python projects categorized as deep-reinforcement-learning | Edit details

Top 23 Python deep-reinforcement-learning Projects

  • GitHub repo trax

    Trax — Deep Learning with Clear Code and Speed

    Project mention: Why would I want to develop yet another deep learning framework? | | 2021-09-16
  • GitHub repo tensorforce

    Tensorforce: a TensorFlow library for applied reinforcement learning

    Project mention: Advice on doing RL for Settlers of Catan? | | 2021-07-11

    The most promising approach has been using the TensorForce framework ( with a custom environment that represents a simpler game (1v1 against a bot that chooses actions randomly, no trading between players, and fixing discarding to be done automatically and at random).

  • Nanos

    Run Linux Software Faster and Safer than Linux with Unikernels.

  • GitHub repo pytorch-a2c-ppo-acktr-gail

    PyTorch implementation of Advantage Actor Critic (A2C), Proximal Policy Optimization (PPO), Scalable trust-region method for deep reinforcement learning using Kronecker-factored approximation (ACKTR) and Generative Adversarial Imitation Learning (GAIL).

    Project mention: How to pretrain a model on expert data? | | 2021-09-12

    Try using an imitation learning algorithm. Two popular options are MaxEnt IRL and GAIL. This repository has GAIL implementation and this repository has MaxEnt IRL and GAIL implementation. There are other implementations too that you can check out.

  • GitHub repo minimalRL

    Implementations of basic RL algorithms with minimal lines of codes! (pytorch based)

    Project mention: Rl algorithm implemented | | 2021-07-18
  • GitHub repo muzero-general


    Project mention: MuZero unable to solve non-slippery FrozenLake environment? | | 2021-08-09

    I have used this implementation from MuZero:

  • GitHub repo rlcard

    Reinforcement Learning / AI Bots in Card (Poker) Games - Blackjack, Leduc, Texas, DouDizhu, Mahjong, UNO.

    Project mention: What sort of algorithm should I use ? Incomplete information, card game. (Flowchart for reference) | | 2021-01-12

    Probably the easiest way for you to get started is to implement your game on an open source RL framework that has working implementations of some basic CFR variations as well as some other self-play algorithms such as NFSP. OpenSpiel and RLCard are two that I am aware of. Depending on the complexity of your game and how strong your agent needs to play, you might be satisfied with the performance you get using by one of these frameworks.

  • GitHub repo ElegantRL

    Lightweight and scalable deep reinforcement learning using PyTorch. 🔥

    Project mention: ElegantRL: A Lightweight and Stable Deep Reinforcement Learning Library | | 2021-03-15
  • Scout APM

    Scout APM: A developer's best friend. Try free for 14-days. Scout APM uses tracing logic that ties bottlenecks to source code so you know the exact line of code causing performance issues and can get back to building a great product faster.

  • GitHub repo deep-q-learning

    Minimal Deep Q Learning (DQN & DDQN) implementations in Keras (by keon)

    Project mention: Deep Q Network knapsack problem | | 2021-05-22

    So go online on GitHub and find a DQN implementation that has options for using a feedforward net as input (instead of conv net as your input isn’t pixel based). Any remotely modular piece of code will take in state space size and action space as parameters to their NN. This is essentially setting input layer to be equal to state space (so 4) and output layer to be action space (201). ( this repo seems helpful i a cursory glance

  • GitHub repo softlearning

    Softlearning is a reinforcement learning framework for training maximum entropy policies in continuous domains. Includes the official implementation of the Soft Actor-Critic algorithm.

    Project mention: DDPG not solving MountainCarContinuous | | 2021-08-30

    You may read - issue with SAC ( ), solution: use large OU noise or use other type of exploration like gSDE

  • GitHub repo habitat-lab

    A modular high-level library to train embodied AI agents across a variety of tasks, environments, and simulators.

    Project mention: Accelerate PPO training | | 2021-07-22
  • GitHub repo deepdrive

    Deepdrive is a simulator that allows anyone with a PC to push the state-of-the-art in self-driving

    Project mention: Is it possible to train a self driving car on google colab? | | 2021-09-01

    I've been trying for a while now and I started thinking it may not be possible. If anyone has managed to train a self-driving car simulator using openai gym on google colab(preferably), or on any remote server (AWS, GCP, ...) please let me know. So far, I tried carla, airsim, svl, deepdrive and they are all equally useless unless run locally with a gui. I'd really appreciate if someone suggests some way that actually can make it possible.

  • GitHub repo ai-economist

    Foundation is a flexible, modular, and composable framework to model socio-economic behaviors and dynamics with both agents and governments. This framework can be used in conjunction with reinforcement learning to learn optimal economic policies, as done by the AI Economist (

    Project mention: "The AI Economist: Optimal Economic Policy Design via Two-level Deep Reinforcement Learning", Zheng et al 2021 {Salesforce} | | 2021-08-05
  • GitHub repo autonomous-learning-library

    A PyTorch library for building deep reinforcement learning agents.

    Project mention: Where do people get their algorithm implementations from? | | 2021-01-27

    I very strongly recommend the autonomous learning library:

  • GitHub repo DeepRL-TensorFlow2

    🐋 Simple implementations of various popular Deep Reinforcement Learning algorithms using TensorFlow2

    Project mention: PPO implementation in TensorFlow2 | | 2021-09-12

    I've been searching for a clean, good, and understandable implementation of PPO for continuous action space with TF2 witch is understandable enough for me to apply my modifications, but the closest thing that I have found is this code which seems to not work properly even on a simple gym cartpole env (discussed issues in git-hub repo suggest the same problem) so I have some doubts :). I was wondering whether you could recommend an implementation that you trust and suggest :)

  • GitHub repo PCGrad

    Code for "Gradient Surgery for Multi-Task Learning"

    Project mention: Help with studying AI in go | | 2021-02-08

    Let me see if I find a public example: the other day I was trying some experiments with PCGrad, so I looked at the code and bumped into this line:

  • GitHub repo drl_grasping

    Deep Reinforcement Learning for Robotic Grasping from Octrees

    Project mention: ROS 2 + Ignition + OpenAI Gym Deep RL Example | | 2021-05-05
  • GitHub repo policy-adaptation-during-deployment

    Training code and evaluation benchmarks for the "Self-Supervised Policy Adaptation during Deployment" paper.

    Project mention: Exploring Self-Supervised Policy Adaptation To Continue Training After Deployment Without Using Any Rewards | | 2021-03-03


  • GitHub repo TensorFlow2.0-for-Deep-Reinforcement-Learning

    TensorFlow 2.0 for Deep Reinforcement Learning. :octopus:

    Project mention: Beginner attempting to implement Noisy DQN | | 2021-01-15

    I forgot to say that I'm using tensorflow, nevertheless I managed to find a git implementation for tensorflow 2 of the noisy dense layer ( and tried to adapt it to my needs.

  • GitHub repo fred

    This my 3d printed robot arm project (by thomashiemstra)

    Project mention: Anyone know of any good robotic arm programming simulators? | | 2021-09-08

    Another option: pybullet. If you just want to learn and you don't need a super high performant system go with python. As an example my code for controlling a simulated robot arm with pybullet my code

  • GitHub repo gym-microrts

    The Reinforcement Learning environment for AI research in μRTS, a Real-time Strategy game simulator.

    Project mention: MARL games in which predicting other agents' behaviour is key to success | | 2021-04-14

    Hey have you looked at

  • GitHub repo neptune-contrib

    Tools, helpers and everything else that helps you work with Neptune.

  • GitHub repo f-IRL

    Inverse Reinforcement Learning via State Marginal Matching - CoRL 2020

    Project mention: How to pretrain a model on expert data? | | 2021-09-12

    Try using an imitation learning algorithm. Two popular options are MaxEnt IRL and GAIL. This repository has GAIL implementation and this repository has MaxEnt IRL and GAIL implementation. There are other implementations too that you can check out.

  • GitHub repo deep_control

    Deep Reinforcement Learning for Continuous Control in PyTorch

    Project mention: Help on what could be wrong on my TD3? | | 2021-06-06

    So I am training with my own simulator from Unity connected to Open AI gym using TD3 adopted from this

NOTE: The open source projects on this list are ordered by number of github stars. The number of mentions indicates repo mentiontions in the last 12 Months or since we started tracking (Dec 2020). The latest post mention was on 2021-09-16.


What are some of the best open-source deep-reinforcement-learning projects in Python? This list will help you:

Project Stars
1 trax 6,467
2 tensorforce 3,018
3 pytorch-a2c-ppo-acktr-gail 2,530
4 minimalRL 1,984
5 muzero-general 1,440
6 rlcard 1,419
7 ElegantRL 1,179
8 deep-q-learning 1,060
9 softlearning 812
10 habitat-lab 702
11 deepdrive 685
12 ai-economist 671
13 autonomous-learning-library 487
14 DeepRL-TensorFlow2 379
15 PCGrad 160
16 drl_grasping 100
17 policy-adaptation-during-deployment 82
18 TensorFlow2.0-for-Deep-Reinforcement-Learning 69
19 fred 63
20 gym-microrts 49
21 neptune-contrib 26
22 f-IRL 21
23 deep_control 12
Find remote jobs at our new job board There are 25 new remote jobs listed recently.
Are you hiring? Post a new remote job listing for free.
SaaSHub - Software Alternatives and Reviews
SaaSHub helps you find the best software and product alternatives