【Spinning up】零、DRLib:一个简洁的强化学习库，集成了HER和PER

六、DRLib:一个简洁的强化学习库，集成了HER和PER

My DRL library with tensorflow1.14 and pytorch, add HER and PER, core codes based on spinning up

前言：

求star，欢迎开issues！
主要的框架，基于spinningup。
目前分享出来的有，tf1和torch版本的主流offpolicy强化算法，DDPG，SAC，TD3，对每个算法都增加了非常简洁的HER和PER的封装，方便调用。
目前仍然是单进程，多进程正在学。
on-policy系列的PPO还没整理好，这周加上去。
离散动作的DQN也有现成的代码，没有封装好。

DRLib：A concise deep reinforcement learning library, integrating HER and PER for almost off policy RL algos

A concise deep reinforcement learning library, integrating HER and PER for almost off policy RL algos.
With tensorflow1.14 and pytorch, add HER and PER, core codes based on https://github.com/openai/spinningup

Compared with spinning up, I delete multi-process and experimental grid wrapper, and our advantage is that it is convenient to debug with pycharm~

项目特点：

1. tf1和pytorch两个版本的算法，前者快，后者新，任君选择；

2. 在spinup的基础上，封装了DDPG, TD3, SAC等主流强化算法，相比原来的函数形式的封装，调用更方便，且加了pytorch的GPU调用；

3. 添加了HER和PER功能，非常适合做机器人相关任务的同学们；

4. 去除了自动调参（ExperimentGrid）和多进程（MPI_fork）部分，适合新手在pycharm中debug，前者直接跑经常会报错~
   等我熟练了这两个，我再加上去，并附上详细教程；

5. 最后，全网最详细的环境配置教程！亲测两个小时内，从零配置完全套环境！

6. 求三连，不行求个star！

1. Installation

1. Clone the repo and cd into it:

   git clone https://github.com/kaixindelele/DRLib.git
   cd DRLib
   

2. Create anaconda DRLib_env env:

   conda create -n DRLib_env python=3.6.9
   source activate DRLib_env
   

3. Install pip_requirement.txt:

   pip install -r pip_requirement.txt
   

   If installation of mpi4py fails, try the following command(Only this one can be installed successfully!):

   conda install mpi4py
   

4. Install tensorflow-gpu=1.14.0

   conda install tensorflow-gpu==1.14.0 # if you have a CUDA-compatible gpu and proper drivers
   

5. Install torch torchvision

   # CUDA 9.2
   conda install pytorch==1.6.0 torchvision==0.7.0 cudatoolkit=9.2 -c pytorch
   
   # CUDA 10.1
   conda install pytorch==1.6.0 torchvision==0.7.0 cudatoolkit=10.1 -c pytorch
   
   # CUDA 10.2
   conda install pytorch==1.6.0 torchvision==0.7.0 cudatoolkit=10.2 -c pytorch
   
   # CPU Only
   conda install pytorch==1.6.0 torchvision==0.7.0 cpuonly -c pytorch
   
   # or pip install    
   pip --default-timeout=100 install torch -i  http://pypi.douban.com/simple  --trusted-host pypi.douban.com
   [pip install torch 在线安装!非离线!](https://blog.csdn.net/hehedadaq/article/details/111480313)
   

6. Install mujoco and mujoco-py

   refer to: https://blog.csdn.net/hehedadaq/article/details/109012048
   

7. Install gym[all]

   refer to https://blog.csdn.net/hehedadaq/article/details/110423154
   

2. Training models

• Example 1. SAC-tf1-HER-PER with FetchPush-v1:

1. modify params in arguments.py, choose env, RL-algorithm, use PER and HER or not, gpu-id, and so on.
2. run with train_tf.py or train_torch.py

   python train_tf.py
   

3. exp results to local:https://blog.csdn.net/hehedadaq/article/details/114045615
4. plot results:https://blog.csdn.net/hehedadaq/article/details/114044217

3. File tree and introduction:

.
├── algos
│   ├── pytorch
│   │   ├── ddpg_sp
│   │   │   ├── core.py-------------It's copied directly from spinup, and modified some details.
│   │   │   ├── ddpg_per_her.py-----inherits from offPolicy.baseOffPolicy, can choose whether or not HER and PER
│   │   │   ├── ddpg.py-------------It's copied directly from spinup
│   │   │   ├── __init__.py
│   │   ├── __init__.py
│   │   ├── offPolicy
│   │   │   ├── baseOffPolicy.py----baseOffPolicy, can be used to DDPG/TD3/SAC and so on.
│   │   │   ├── norm.py-------------state normalizer, update mean/std with training process.
│   │   ├── sac_auto
│   │   ├── sac_sp
│   │   │   ├── core.py-------------likely as before.
│   │   │   ├── __init__.py
│   │   │   ├── sac_per_her.py
│   │   │   └── sac.py
│   │   └── td3_sp
│   │       ├── core.py
│   │       ├── __init__.py
│   │       ├── td3_gpu_class.py----td3_class modified from spinup
│   │       └── td3_per_her.py
│   └── tf1
│       ├── ddpg_sp
│       │   ├── core.py
│       │   ├── DDPG_class.py------------It's copied directly from spinup, and wrap algorithm from function to class.
│       │   ├── DDPG_per_class.py--------Add PER.
│       │   ├── DDPG_per_her_class.py----DDPG with HER and PER without inheriting from offPolicy.
│       │   ├── DDPG_per_her.py----------Add HER and PER.
│       │   ├── DDPG_sp.py---------------It's copied directly from spinup, and modified some details.
│       │   ├── __init__.py
│       ├── __init__.py
│       ├── offPolicy
│       │   ├── baseOffPolicy.py
│       │   ├── core.py
│       │   ├── norm.py
│       ├── sac_auto--------------------SAC with auto adjust alpha parameter version.
│       │   ├── core.py
│       │   ├── __init__.py
│       │   ├── sac_auto_class.py
│       │   ├── sac_auto_per_class.py
│       │   └── sac_auto_per_her.py
│       ├── sac_sp--------------------SAC with alpha=0.2 version.
│       │   ├── core.py
│       │   ├── __init__.py
│       │   ├── SAC_class.py
│       │   ├── SAC_per_class.py
│       │   ├── SAC_per_her.py
│       │   ├── SAC_sp.py
│       └── td3_sp
│           ├── core.py
│           ├── __init__.py
│           ├── TD3_class.py
│           ├── TD3_per_class.py
│           ├── TD3_per_her_class.py
│           ├── TD3_per_her.py
│           ├── TD3_sp.py
├── arguments.py-----------------------hyperparams scripts
├── drlib_tree.txt
├── HER_DRLib_exps---------------------demo exp logs
│   ├── 2021-02-21_HER_TD3_FetchPush-v1
│   │   ├── 2021-02-21_18-26-08-HER_TD3_FetchPush-v1_s123
│   │   │   ├── checkpoint
│   │   │   ├── config.json
│   │   │   ├── params.data-00000-of-00001
│   │   │   ├── params.index
│   │   │   ├── progress.txt
│   │   │   └── Script_backup.py
├── memory
│   ├── __init__.py
│   ├── per_memory.py--------------mofan version
│   ├── simple_memory.py-----------mofan version
│   ├── sp_memory.py---------------spinningup tf1 version, simple uniform buffer memory class.
│   ├── sp_memory_torch.py---------spinningup torch-gpu version, simple uniform buffer memory class.
│   ├── sp_per_memory.py-----------spinningup tf1 version, PER buffer memory class.
│   └── sp_per_memory_torch.py
├── pip_requirement.txt------------pip install requirement, exclude mujoco-py,gym,tf,torch.
├── spinup_utils-------------------some utils from spinningup, about ploting results, logging, and so on.
│   ├── delete_no_checkpoint.py----delete the folder where the experiment did not complete.
│   ├── __init__.py
│   ├── logx.py
│   ├── mpi_tf.py
│   ├── mpi_tools.py
│   ├── plot.py
│   ├── print_logger.py------------save the information printed by the terminal to the local log file。
│   ├── run_utils.py---------------now I haven't used it. I have to learn how to multi-process.
│   ├── serialization_utils.py
│   └── user_config.py
├── train_tf1.py--------------main.py for tf1
└── train_torch.py------------main.py for torch

4. HER introduction:

Refer to these code bases:

1. It can be converged, but this code is too difficult. https://github.com/openai/baselines

2. It can also converged, but only for DDPG-torch-cpu. https://github.com/sush1996/DDPG_Fetch

3. It can not be converged, but this code is simpler. https://github.com/Stable-Baselines-Team/stable-baselines

• paper: https://arxiv.org/pdf/1709.10089.pdf

4.1. My understanding and video:

种瓜得豆来解释her:
第一步在春天（state），种瓜（origin-goal）得豆，通过HER，把目标换成种豆，按照之前的操作，可以学会在春天种豆得豆；
第二步种米得瓜，学会种瓜得瓜；
即只要是智能体中间经历过的状态，都可以当做它的目标，进行学会。
即如果智能体能遍历所有的状态空间，那么它就可以学会达到整个状态空间。

https://www.bilibili.com/video/BV1BA411x7Wm

4.2. Key tricks for HER:

1. state-normalize: success rate from 0 to 1 for FetchPush-v1 task.
2. Q-clip: success rate from 0.5 to 0.7 for FetchPickAndPlace-v1 task.
3. action_l2: little effect for Push task.

4.3. Performance about HER-DDPG with FetchPush-v1:

5. PER introduction:

refer to:off-policy全系列（DDPG-TD3-SAC-SAC-auto）+优先经验回放PER-代码-实验结果分析

6. Summary：

这个库我封装了好久，整个代码库简洁、方便、功能比较齐全，在环境配置这块几乎是手把手教程，希望能给大家节省一些时间~

从零开始配置，不到两小时，从下载代码库，到配置环境，到在自己的环境中跑通，全流程非常流畅。

6.1. 下一步添加的功能：

1. PPO的封装；

2. DQN的封装；

3. 多进程的封装；

4. ExperimentGrid的封装；

7. Contact：

深度强化学习-DRL：799378128

欢迎关注知乎帐号：未入门的炼丹学徒

CSDN帐号：https://blog.csdn.net/hehedadaq