Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/StepNeverStop/RLs
Reinforcement Learning Algorithms Based on PyTorch
https://github.com/StepNeverStop/RLs
deep-reinforcement-learning gym ml-agents pytorch reinforcement-learning reinforcement-learning-algorithm sac training-agents unity3d
Last synced: about 2 months ago
JSON representation
Reinforcement Learning Algorithms Based on PyTorch
- Host: GitHub
- URL: https://github.com/StepNeverStop/RLs
- Owner: StepNeverStop
- License: apache-2.0
- Created: 2019-04-25T16:44:54.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2021-10-21T15:23:48.000Z (almost 3 years ago)
- Last Synced: 2024-07-29T06:37:49.276Z (about 2 months ago)
- Topics: deep-reinforcement-learning, gym, ml-agents, pytorch, reinforcement-learning, reinforcement-learning-algorithm, sac, training-agents, unity3d
- Language: Python
- Homepage: https://stepneverstop.github.io
- Size: 11.7 MB
- Stars: 446
- Watchers: 18
- Forks: 96
- Open Issues: 24
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
RLs: Reinforcement Learning Algorithm Based On PyTorch.
# RLs
This project includes SOTA or classic reinforcement learning (single and multi-agent) algorithms used for training
agents by interacting with Unity through [ml-agents](https://github.com/Unity-Technologies/ml-agents/tree/release_18)
Release 18 or with [gym](https://github.com/openai/gym).
## About
The goal of this framework is to provide stable implementations of standard RL algorithms and simultaneously enable fast
prototyping of new methods. It aims to fill the need for a small, easily grokked codebase in which users can freely
experiment with wild ideas (speculative research).
## Characteristics
This project supports:
- Suitable for Windows, Linux, and OSX
- Single- and Multi-Agent training.
- Multiple type of observation sensors as input.
- Only need 3 steps to implement a new algorithm:
1. **policy** write `.py` in `rls/algorithms/{single/multi}` directory and make the policy inherit from super-class
defined in `rls/algorithms/base`
2. **config** write `.yaml` in `rls/configs/algorithms/` directory and specify the super config type defined
in `rls/configs/algorithms/general.yaml`
3. **register** register new algorithm in `rls/algorithms/__init__.py`
- Only need 3 steps to adapt to a new training environment:
1. **wrapper** write environment wrappers in `rls/envs/{new platform}` directory and make it inherit from
super-class defined in `rls/envs/env_base.py`
2. **config** write default configuration in `rls/configs/{new platform}`
3. **register** register new environment platform in `rls/envs/__init__.py`
- Compatible with several environment platforms
- [Unity3D ml-agents](https://github.com/Unity-Technologies/ml-agents).
- [PettingZoo](https://www.pettingzoo.ml/#)
- [gym](https://github.com/openai/gym), for now only two data types are compatible——`[Box, Discrete]`. Support
parallel training using gym envs, just need to specify `--copies` to how many agents you want to train in
parallel.
- environments:
- [MuJoCo](https://github.com/openai/mujoco-py)(v2.0.2.13)
- [PyBullet](https://github.com/bulletphysics/bullet3)
- [gym_minigrid](https://github.com/maximecb/gym-minigrid)
- observation -> action:
- Discrete -> Discrete (observation type -> action type)
- Discrete -> Box
- Box -> Discrete
- Box -> Box
- Box/Discrete -> Tuple(Discrete, Discrete, Discrete)
- Four types of Replay Buffer, Default is ER:
- ER
- [Prioritized ER](https://arxiv.org/abs/1511.05952)
- [Noisy Net](https://arxiv.org/abs/1706.10295) for better exploration.
- [Intrinsic Curiosity Module](https://arxiv.org/abs/1705.05363) for almost all off-policy algorithms implemented.
- Parallel training multiple scenes for Gym
- Unified data format
## Installation
method 1:
```bash
$ git clone https://github.com/StepNeverStop/RLs.git
$ cd RLs
$ conda create -n rls python=3.8
$ conda activate rls
# Windows
$ pip install -e .[windows]
# Linux or Mac OS
$ pip install -e .
```
method 1:
```bash
conda env create -f environment.yaml
```
If using ml-agents:
```bash
$ pip install -e .[unity]
```
You can download the builded docker image from [here](https://hub.docker.com/r/keavnn/rls):
```bash
$ docker pull keavnn/rls:latest
```
If anyone who wants to send a PR, plz format all code-files first:
```bash
$ pip install -e .[pr]
$ python auto_format.py -d ./
```
## Implemented Algorithms
For now, these algorithms are available:
- Multi-Agent training algorithms:
- Independent-SARL, i.e. IQL, [I-DQN](http://arxiv.org/abs/1511.08779), etc.
- [Value-Decomposition Networks, VDN](http://arxiv.org/abs/1706.05296)
- [Monotonic Value Function Factorisation Networks, QMIX](http://arxiv.org/abs/1803.11485)
- [Multi-head Attention based Q-value Mixing Network, Qatten](http://arxiv.org/abs/2002.03939)
- [Factorize with Transformation, Qtran](https://arxiv.org/abs/1905.05408)
- [Duplex Dueling Multi-Agent Q-Learning, QPLEX](http://arxiv.org/abs/2008.01062)
- [Multi-Agent Deep Deterministic Policy Gradient, MADDPG](https://arxiv.org/abs/1706.02275)
- Single-Agent training algorithms(Some algorithms that only support continuous space problems use Gumbel-softmax trick
to implement discrete versions, i.e. DDPG):
- Policy Gradient, PG
- Actor Critic, AC
- [Synchronous Advantage Actor Critic, A2C](http://arxiv.org/abs/1602.01783)
- :boom:Proximal Policy Optimization, [PPO](https://arxiv.org/abs/1707.06347)
, [DPPO](http://arxiv.org/abs/1707.02286,)
- [Trust Region Policy Optimization, TRPO](https://arxiv.org/abs/1502.05477)
- [Natural Policy Gradient, NPG](https://proceedings.neurips.cc/paper/2001/file/4b86abe48d358ecf194c56c69108433e-Paper.pdf)
- [Deterministic Policy Gradient, DPG](https://hal.inria.fr/file/index/docid/938992/filename/dpg-icml2014.pdf)
- [Deep Deterministic Policy Gradient, DDPG](https://arxiv.org/abs/1509.02971)
- :fire:Soft Actor Critic, [SAC](https://arxiv.org/abs/1812.05905), [Discrete SAC](https://arxiv.org/abs/1910.07207)
- [Tsallis Actor Critic, TAC](https://arxiv.org/abs/1902.00137)
- :fire:[Twin Delayed Deep Deterministic Policy Gradient, TD3](https://arxiv.org/abs/1802.09477)
- Deep Q-learning Network, DQN, [2013](https://arxiv.org/pdf/1312.5602.pdf)
, [2015](https://storage.googleapis.com/deepmind-media/dqn/DQNNaturePaper.pdf)
- [Double Deep Q-learning Network, DDQN](https://arxiv.org/abs/1509.06461)
- [Dueling Double Deep Q-learning Network, DDDQN](https://arxiv.org/abs/1511.06581)
- [Deep Recurrent Q-learning Network, DRQN](https://arxiv.org/abs/1507.06527)
- [Deep Recurrent Double Q-learning, DRDQN](https://arxiv.org/abs/1908.06040)
- [Category 51, C51](https://arxiv.org/abs/1707.06887)
- [Quantile Regression DQN, QR-DQN](https://arxiv.org/abs/1710.10044)
- [Implicit Quantile Networks, IQN](https://arxiv.org/abs/1806.06923)
- [Rainbow DQN](https://arxiv.org/abs/1710.02298)
- [MaxSQN](https://github.com/createamind/DRL/blob/master/spinup/algos/maxsqn/maxsqn.py)
- [Soft Q-Learning, SQL](https://arxiv.org/abs/1702.08165)
- [Bootstrapped DQN](http://arxiv.org/abs/1602.04621)
- [Averaged DQN](http://arxiv.org/abs/1611.01929)
- Hierachical training algorithms:
- [Option-Critic, OC](http://arxiv.org/abs/1609.05140)
- [Asynchronous Advantage Option-Critic, A2OC](http://arxiv.org/abs/1709.04571)
- [PPO Option-Critic, PPOC](http://arxiv.org/abs/1712.00004)
- [Interest-Option-Critic, IOC](http://arxiv.org/abs/2001.00271)
- Model-based algorithms:
- [Learning Latent Dynamics for Planning from Pixels, PlaNet](http://arxiv.org/abs/1811.04551)
- [Dream to Control, Dreamer](http://arxiv.org/abs/1912.01603)
- [Mastering Atari with Discrete World Models, DreamerV2](http://arxiv.org/abs/2010.02193)
- [Model-Based Value Estimation, MVE](http://arxiv.org/abs/1803.00101)
- Offline algorithms(**under implementation**):
- [Conservative Q-Learning for Offline Reinforcement Learning, CQL](http://arxiv.org/abs/2006.04779)
- BCQ
- Benchmarking Batch Deep Reinforcement Learning Algorithms, [Discrete](http://arxiv.org/abs/1910.01708)
- Off-Policy Deep Reinforcement Learning without Exploration, [Continuous](http://arxiv.org/abs/1812.02900)
| Algorithms | Discrete | Continuous | Image | RNN | Command parameter |
| :-----------------------------: | :------: | :--------: | :---: | :--: | :---------------: |
| PG | ✓ | ✓ | ✓ | ✓ | pg |
| AC | ✓ | ✓ | ✓ | ✓ | ac |
| A2C | ✓ | ✓ | ✓ | ✓ | a2c |
| NPG | ✓ | ✓ | ✓ | ✓ | npg |
| TRPO | ✓ | ✓ | ✓ | ✓ | trpo |
| PPO | ✓ | ✓ | ✓ | ✓ | ppo |
| DQN | ✓ | | ✓ | ✓ | dqn |
| Double DQN | ✓ | | ✓ | ✓ | ddqn |
| Dueling Double DQN | ✓ | | ✓ | ✓ | dddqn |
| Averaged DQN | ✓ | | ✓ | ✓ | averaged_dqn |
| Bootstrapped DQN | ✓ | | ✓ | ✓ | bootstrappeddqn |
| Soft Q-Learning | ✓ | | ✓ | ✓ | sql |
| C51 | ✓ | | ✓ | ✓ | c51 |
| QR-DQN | ✓ | | ✓ | ✓ | qrdqn |
| IQN | ✓ | | ✓ | ✓ | iqn |
| Rainbow | ✓ | | ✓ | ✓ | rainbow |
| DPG | ✓ | ✓ | ✓ | ✓ | dpg |
| DDPG | ✓ | ✓ | ✓ | ✓ | ddpg |
| TD3 | ✓ | ✓ | ✓ | ✓ | td3 |
| SAC(has V network) | ✓ | ✓ | ✓ | ✓ | sac_v |
| SAC | ✓ | ✓ | ✓ | ✓ | sac |
| TAC | sac | ✓ | ✓ | ✓ | tac |
| MaxSQN | ✓ | | ✓ | ✓ | maxsqn |
| OC | ✓ | ✓ | ✓ | ✓ | oc |
| AOC | ✓ | ✓ | ✓ | ✓ | aoc |
| PPOC | ✓ | ✓ | ✓ | ✓ | ppoc |
| IOC | ✓ | ✓ | ✓ | ✓ | ioc |
| PlaNet | ✓ | | ✓ | 1 | planet |
| Dreamer | ✓ | ✓ | ✓ | 1 | dreamer |
| DreamerV2 | ✓ | ✓ | ✓ | 1 | dreamerv2 |
| VDN | ✓ | | ✓ | ✓ | vdn |
| QMIX | ✓ | | ✓ | ✓ | qmix |
| Qatten | ✓ | | ✓ | ✓ | qatten |
| QPLEX | ✓ | | ✓ | ✓ | qplex |
| QTRAN | ✓ | | ✓ | ✓ | qtran |
| MADDPG | ✓ | ✓ | ✓ | ✓ | maddpg |
| MASAC | ✓ | ✓ | ✓ | ✓ | masac |
| CQL | ✓ | | ✓ | ✓ | cql_dqn |
| BCQ | ✓ | ✓ | ✓ | ✓ | bcq |
| MVE | ✓ | ✓ | | | mve |
*1 means must use rnn or rnn is used by default.*
## Getting started
```python
"""
usage: run.py [-h] [-c COPIES] [--seed SEED] [-r]
[-p {gym,unity,pettingzoo}]
[-a {maddpg,masac,vdn,qmix,qatten,qtran,qplex,aoc,ppoc,oc,ioc,planet,dreamer,dreamerv2,mve,cql_dqn,bcq,pg,npg,trpo,ppo,a2c,ac,dpg,ddpg,td3,sac_v,sac,tac,dqn,ddqn,dddqn,averaged_dqn,c51,qrdqn,rainbow,iqn,maxsqn,sql,bootstrappeddqn}]
[-i] [-l LOAD_PATH] [-m MODELS] [-n NAME]
[--config-file CONFIG_FILE] [--store-dir STORE_DIR]
[--episode-length EPISODE_LENGTH] [--hostname] [-e ENV_NAME]
[-f FILE_NAME] [-s] [-d DEVICE] [-t MAX_TRAIN_STEP]
optional arguments:
-h, --help show this help message and exit
-c COPIES, --copies COPIES
nums of environment copies that collect data in
parallel
--seed SEED specify the random seed of module random, numpy and
pytorch
-r, --render whether render game interface
-p {gym,unity,pettingzoo}, --platform {gym,unity,pettingzoo}
specify the platform of training environment
-a {maddpg,masac,vdn,qmix,qatten,qtran,qplex,aoc,ppoc,oc,ioc,planet,dreamer,dreamerv2,mve,cql_dqn,bcq,pg,npg,trpo,ppo,a2c,ac,dpg,ddpg,td3,sac_v,sac,tac,dqn,ddqn,dddqn,averaged_dqn,c51,qrdqn,rainbow,iqn,maxsqn,sql,bootstrappeddqn}, --algorithm {maddpg,masac,vdn,qmix,qatten,qtran,qplex,aoc,ppoc,oc,ioc,planet,dreamer,dreamerv2,mve,cql_dqn,bcq,pg,npg,trpo,ppo,a2c,ac,dpg,ddpg,td3,sac_v,sac,tac,dqn,ddqn,dddqn,averaged_dqn,c51,qrdqn,rainbow,iqn,maxsqn,sql,bootstrappeddqn}
specify the training algorithm
-i, --inference inference the trained model, not train policies
-l LOAD_PATH, --load-path LOAD_PATH
specify the name of pre-trained model that need to
load
-m MODELS, --models MODELS
specify the number of trails that using different
random seeds
-n NAME, --name NAME specify the name of this training task
--config-file CONFIG_FILE
specify the path of training configuration file
--store-dir STORE_DIR
specify the directory that store model, log and
others
--episode-length EPISODE_LENGTH
specify the maximum step per episode
--hostname whether concatenate hostname with the training name
-e ENV_NAME, --env-name ENV_NAME
specify the environment name
-f FILE_NAME, --file-name FILE_NAME
specify the path of builded training environment of
UNITY3D
-s, --save specify whether save models/logs/summaries while
training or not
-d DEVICE, --device DEVICE
specify the device that operate Torch.Tensor
-t MAX_TRAIN_STEP, --max-train-step MAX_TRAIN_STEP
specify the maximum training steps
"""
```
Example:
```bash
python run.py -s # save model and log while train
python run.py -p gym -a dqn -e CartPole-v0 -c 12 -n dqn_cartpole
python run.py -p unity -a ppo -n run_with_unity -c 1
```
The main training loop of **pseudo-code** in this repo is as:
```python
# noinspection PyUnresolvedReferences
agent.episode_reset() # initialize rnn hidden state or something else
# noinspection PyUnresolvedReferences
obs = env.reset()
while True:
# noinspection PyUnresolvedReferences
env_rets = env.step(agent(obs))
# noinspection PyUnresolvedReferences
agent.episode_step(obs, env_rets) # store experience, save model, and train off-policy algorithms
obs = env_rets['obs']
if env_rets['done']:
break
# noinspection PyUnresolvedReferences
agent.episode_end() # train on-policy algorithms
```
## Giving credit
If using this repository for your research, please cite:
```
@misc{RLs,
author = {Keavnn},
title = {RLs: A Featureless Reinforcement Learning Repository},
year = {2019},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/StepNeverStop/RLs}},
}
```
## Issues
Any questions/errors about this project, please let me know in [here](https://github.com/StepNeverStop/RLs/issues/new).