深度强化学习环境几次在不同环境中迁移，特此写博客以便记录下次配置安装。

Anaconda 3

wget https://mirrors.tuna.tsinghua.edu.cn/anaconda/archive/Anaconda3-5.1.0-Linux-x86_64.sh
bash Anaconda3-5.1.0-Linux-x86.sh

之后在 运行 bashrc更改 source .bashrc
该安装包会自动安装所有包，加环境变量，以及安装VSCODE。

conda 导入environment.xml

name: pytorch1.0
channels:
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/msys2/
  - https://mirrors.ustc.edu.cn/anaconda/pkgs/free/
  - defaults
dependencies:
  - ca-certificates=2019.1.23=0
  - certifi=2018.11.29=py36_0
  - libedit=3.1.20181209=hc058e9b_0
  - libffi=3.2.1=hd88cf55_4
  - libgcc-ng=8.2.0=hdf63c60_1
  - libstdcxx-ng=8.2.0=hdf63c60_1
  - ncurses=6.1=he6710b0_1
  - openssl=1.0.2r=h7b6447c_0
  - pip=19.0.3=py36_0
  - python=3.6.5=hc3d631a_2
  - readline=7.0=h7b6447c_5
  - setuptools=40.8.0=py36_0
  - sqlite=3.26.0=h7b6447c_0
  - tk=8.6.8=hbc83047_0
  - wheel=0.33.1=py36_0
  - xz=5.2.4=h14c3975_4
  - zlib=1.2.11=h7b6447c_3
  - pip:
    - atari-py==0.1.7
    - atomicwrites==1.3.0
    - attrs==19.1.0
    - cffi==1.12.3
    - chardet==3.0.4
    - click==7.0
    - colorama==0.4.1
    - cycler==0.10.0
    - cython==0.29.13
    - decorator==4.4.0
    - docopt==0.6.2
    - filelock==3.0.12
    - flask==1.1.1
    - flatbuffers==1.11
    - funcsigs==1.0.2
    - future==0.17.1
    - glfw==1.8.2
    - gym==0.12.1
    - idna==2.8
    - imageio==2.5.0
    - importlib-metadata==0.18
    - itsdangerous==1.1.0
    - jinja2==2.10.1
    - kiwisolver==1.1.0
    - lockfile==0.12.2
    - markupsafe==1.1.1
    - matplotlib==3.1.1
    - more-itertools==7.1.0
    - mujoco-py==2.0.2.4
    - networkx==2.3
    - numpy==1.16.3
    - opencv-python==4.1.1.26
    - packaging==19.0
    - pandas==0.24.2
    - pillow==6.0.0
    - pluggy==0.12.0
    - pqi==2.0.6
    - protobuf==3.9.1
    - psutil==5.6.3
    - py==1.8.0
    - py-spy==0.1.11
    - pycparser==2.19
    - pyglet==1.3.2
    - pyopengl==3.1.0
    - pyparsing==2.4.0
    - pytest==5.0.0
    - python-dateutil==2.8.0
    - pytz==2019.1
    - pywavelets==1.0.3
    - pyyaml==5.1.1
    - ray==0.7.3
    - redis==3.2.1
    - requests==2.21.0
    - scikit-image==0.15.0
    - scipy==1.2.1
    - selectivesearch==0.4
    - setproctitle==1.1.10
    - six==1.12.0
    - torch==1.0.1.post2
    - torchvision==0.2.2.post3
    - typing==3.7.4
    - urllib3==1.24.2
    - wcwidth==0.1.7
    - werkzeug==0.15.5
    - zipp==0.5.1

分布式安装库horovod

一个支持tensorflow,pytorch,keras的分布式库，基于mpirun,github地址https://github.com/horovod/horovod

$ git clone --recursive https://github.com/horovod/horovod
$ cd horovod
$ python setup.py sdist
$ HOROVOD_GPU_ALLREDUCE=NCCL HOROVOD_WITH_TENSORFLOW=1 /data6/yizhou/tool/miniconda3/bin/pip install --no-cache-dir

视频截图render库解决