DI-engine: A Python repository from huang-fuxian

Updated on 2022.01.01 DI-engine-v0.2.3 (beta)

Introduction to DI-engine (beta)

DI-engine is a generalized decision intelligence engine. It supports various deep reinforcement learning algorithms (link):

Most basic DRL algorithms, such as DQN, PPO, SAC, R2D2
Multi-agent RL algorithms like QMIX, MAPPO
Imitation learning algorithms (BC/IRL/GAIL) , such as GAIL, SQIL, Guided Cost Learning
Exploration algorithms like HER, RND, ICM
Offline RL algorithms: CQL, TD3BC
Model-based RL algorithms: MBPO

DI-engine aims to standardize different RL enviroments and applications. Various training pipelines and customized decision AI applications are also supported.

Traditional academic environments
- DI-zoo
Real world decision AI applications
- DI-star: Decision AI in StarCraftII
- DI-drive: Auto-driving platform
- GoBigger: Multi-Agent Decision Intelligence Environment
- DI-smartcross: Decision AI in Traffic Light Control
General nested data lib
- treevalue: Tree-nested data structure
- DI-treetensor: Tree-nested PyTorch tensor Lib
Docs and Tutorials
- DI-engine-docs
- awesome-model-based-RL: A curated list of awesome Model-Based RL resources

DI-engine also has some system optimization and design for efficient and robust large-scale RL training:

DI-orchestrator: RL Kubernetes Custom Resource and Operator Lib
DI-hpc: RL HPC OP Lib
DI-store: RL Object Store

Have fun with exploration and exploitation.

Installation

You can simply install DI-engine from PyPI with the following command:

pip install DI-engine

If you use Anaconda or Miniconda, you can install DI-engine from conda-forge through the following command:

conda install -c opendilab di-engine

For more information about installation, you can refer to installation.

And our dockerhub repo can be found here，we prepare base image and env image with common RL environments.

base: opendilab/ding:nightly
atari: opendilab/ding:nightly-atari
mujoco: opendilab/ding:nightly-mujoco
smac: opendilab/ding:nightly-smac

The detailed documentation are hosted on doc | 中文文档.

Quick Start

3 Minutes Kickoff

3 Minutes Kickoff (colab)

3 分钟上手中文版 (kaggle)

How to migrate a new RL Env | 如何迁移一个新的强化学习环境

Bonus: Train RL agent in one line code:

ding -m serial -e cartpole -p dqn -s 0

Supporters

↳ Stargazers

↳ Forkers

Feature

Algorithm Versatility

No	Algorithm	Doc and Implementation	Runnable Demo
1	DQN	DQN doc DQN中文文档 policy/dqn	python3 -u cartpole_dqn_main.py / ding -m serial -c cartpole_dqn_config.py -s 0
2	C51	policy/c51	ding -m serial -c cartpole_c51_config.py -s 0
3	QRDQN	policy/qrdqn	ding -m serial -c cartpole_qrdqn_config.py -s 0
4	IQN	policy/iqn	ding -m serial -c cartpole_iqn_config.py -s 0
5	Rainbow	policy/rainbow	ding -m serial -c cartpole_rainbow_config.py -s 0
6	SQL	policy/sql	ding -m serial -c cartpole_sql_config.py -s 0
7	R2D2	policy/r2d2	ding -m serial -c cartpole_r2d2_config.py -s 0
8	A2C	policy/a2c	ding -m serial -c cartpole_a2c_config.py -s 0
9	PPO/MAPPO	policy/ppo	python3 -u cartpole_ppo_main.py / ding -m serial_onpolicy -c cartpole_ppo_config.py -s 0
10	PPG	policy/ppg	python3 -u cartpole_ppg_main.py
11	ACER	policy/acer	ding -m serial -c cartpole_acer_config.py -s 0
12	IMPALA	policy/impala	ding -m serial -c cartpole_impala_config.py -s 0
13	DDPG/PADDPG	policy/ddpg	ding -m serial -c pendulum_ddpg_config.py -s 0
14	TD3	policy/td3	python3 -u pendulum_td3_main.py / ding -m serial -c pendulum_td3_config.py -s 0
15	D4PG	policy/d4pg	python3 -u pendulum_d4pg_config.py
16	SAC	policy/sac	ding -m serial -c pendulum_sac_config.py -s 0
17	PDQN	policy/pdqn	ding -m serial -c gym_hybrid_pdqn_config.py -s 0
18	MPDQN	policy/pdqn	ding -m serial -c gym_hybrid_mpdqn_config.py -s 0
19	QMIX	policy/qmix	ding -m serial -c smac_3s5z_qmix_config.py -s 0
20	COMA	policy/coma	ding -m serial -c smac_3s5z_coma_config.py -s 0
21	QTran	policy/qtran	ding -m serial -c smac_3s5z_qtran_config.py -s 0
22	WQMIX	policy/wqmix	ding -m serial -c smac_3s5z_wqmix_config.py -s 0
23	CollaQ	policy/collaq	ding -m serial -c smac_3s5z_collaq_config.py -s 0
24	GAIL	reward_model/gail	ding -m serial_gail -c cartpole_dqn_gail_config.py -s 0
25	SQIL	entry/sqil	ding -m serial_sqil -c cartpole_sqil_config.py -s 0
26	DQFD	policy/dqfd	ding -m serial_dqfd -c cartpole_dqfd_config.py -s 0
27	R2D3	R2D3中文文档 policy/r2d3	python3 -u pong_r2d3_r2d2expert_config.py
28	Guided Cost Learning	reward_model/guided_cost	python3 lunarlander_gcl_config.py
29	TREX	reward_model/trex	python3 mujoco_trex_main.py
30	HER	reward_model/her	python3 -u bitflip_her_dqn.py
31	RND	reward_model/rnd	python3 -u cartpole_ppo_rnd_main.py
32	ICM	ICM中文文档 reward_model/icm	python3 -u cartpole_ppo_icm_config.py
33	CQL	policy/cql	python3 -u d4rl_cql_main.py
34	TD3BC	policy/td3_bc	python3 -u mujoco_td3_bc_main.py
35	MBPO	model/template/model_based/mbpo	python3 -u sac_halfcheetah_mopo_default_config.py
36	PER	worker/replay_buffer	`rainbow demo`
37	GAE	rl_utils/gae	`ppo demo`

means discrete action space, which is only label in normal DRL algorithms (1-18)

means continuous action space, which is only label in normal DRL algorithms (1-18)

means hybrid (discrete + continuous) action space (1-18)

means distributed training (collector-learner parallel) RL algorithm

means multi-agent RL algorithm

means RL algorithm which is related to exploration and sparse reward

means Imitation Learning, including Behaviour Cloning, Inverse RL, Adversarial Structured IL

means offline RL algorithm

means model-based RL algorithm

means other sub-direction algorithm, usually as plugin-in in the whole pipeline

P.S: The .py file in Runnable Demo can be found in dizoo

Environment Versatility

No	Environment	Code and Doc Links
1	atari	code link env tutorial 环境指南
2	box2d/bipedalwalker	dizoo link 环境指南
3	box2d/lunarlander	dizoo link 环境指南
4	classic_control/cartpole	dizoo link 环境指南
5	classic_control/pendulum	dizoo link 环境指南
6	competitive_rl	dizoo link 环境指南
7	gfootball	dizoo link 环境指南
8	minigrid	dizoo link 环境指南
9	mujoco	dizoo link 环境指南
10	multiagent_particle	dizoo link 环境指南
11	overcooked	dizoo link env tutorial
12	procgen	dizoo link 环境指南
13	pybullet	dizoo link 环境指南
14	smac	dizoo link 环境指南
15	d4rl	dizoo link 环境指南
16	league_demo	dizoo link
17	pomdp atari	dizoo link
18	bsuite	dizoo link env tutorial
19	ImageNet	dizoo link 环境指南
20	slime_volleyball	dizoo link env tutorial 环境指南
21	gym_hybrid	dizoo link 环境指南
22	GoBigger	opendilab link env tutorial 环境指南
23	gym_soccer	dizoo link 环境指南
24	multiagent_mujoco	dizoo link 环境指南
25	bitflip	dizoo link 环境指南

means discrete action space

means continuous action space

means hybrid (discrete + continuous) action space

means multi-agent RL environment

means environment which is related to exploration and sparse reward

means offline RL environment

means Imitation Learning or Supervised Learning Dataset

means environment that allows agent VS agent battle

P.S. some enviroments in Atari, such as MontezumaRevenge, are also sparse reward type

Feedback and Contribution

File an issue on Github
Open or participate in our forum
Discuss on DI-engine slack communication channel or QQ group (700157520)
Contributes to our future plan Roadmap

We appreciate all the feedbacks and contributions to improve DI-engine, both algorithms and system designs. And CONTRIBUTING.md offers some necessary information.

Citation

@misc{ding,
    title={{DI-engine: OpenDILab} Decision Intelligence Engine},
    author={DI-engine Contributors},
    publisher = {GitHub},
    howpublished = {\url{https://github.com/opendilab/DI-engine}},
    year={2021},
}

License

DI-engine released under the Apache 2.0 license.

huang-fuxian/DI-engine