automate building conda rllibpack.tar.gz

[slinlee]

Script from dh

1  apt-get update
apt-get install build-essential libcurl4-gnutls-dev libxml2-dev libssl-dev wget -y
history
wget https://repo.anaconda.com/archive/Anaconda3-2019.03-Linux-x86_64.sh
bash Anaconda3-2019.03-Linux-x86_64.sh 
source /root/anaconda3/bin/activate
conda install pybind11
pip install --upgrade pip
pip install tensorflow==2.3.1
pip install mock==4.03
pip install mock==4.0.3
pip install spyder==3.3.3
pip install fire==0.3.1
pip install ray[rllib]==1.3.0
pip install or-gym==0.2.0
pip install aioredis==1.3.1
pip install pathmind
conda install -c conda-forge conda-pack
conda pack -o rllibpack.tar.gz
conda list
conda pack -o rllibpack.tar.gz
mv rllibpack.tar.gz /.pathmind/

// conda
docker run --name conda \
    -it -v $HOME/.pathmind/share:/.pathmind \
    ubuntu:20.04

[slinlee]

the script has been updated to use versions that work with our training

apt-get update
apt-get install build-essential libcurl4-gnutls-dev libxml2-dev libssl-dev wget -y
history
wget https://repo.anaconda.com/archive/Anaconda3-2019.03-Linux-x86_64.sh
bash Anaconda3-2019.03-Linux-x86_64.sh 
source /root/anaconda3/bin/activate
conda install pybind11
pip install --upgrade pip
pip install tensorflow==2.3.1
pip install mock==4.0.3
pip install spyder==3.3.3
pip install fire==0.3.1
pip install ray[rllib]==1.3.0
pip install or-gym==0.2.0
pip install aioredis==1.3.1
pip install pathmind
pip install gym==0.18.0
pip install numpy==1.16.2
conda install -c conda-forge conda-pack
conda pack -o rllibpack.tar.gz
conda list
conda pack -o rllibpack.tar.gz
mv rllibpack.tar.gz /.pathmind/