gabbage / transfer_rewards

A Reward Training Framework

Note

Your interest to this project is highly appreciated. Please give a Github star (on top right) if you use the code in anyhow.

Setup

• install anaconda (my anaconda version:2018.12, python version : 3.7.1.final.0, conda version:4.5.12, conda-build version : 3.17.6) (download it from its website). You can check this information by the following commands:

conda list conda$
conda info

• create a virtual environment by anaconda

conda env create -f ./documentation/environment-gpu.yml

Run

If you run the project for the first time, you should update the machine variable in the run.sh based on the type of the machine.

In order to run the agent:

./run.sh

If you want to finish the conversation, type bye

You can check the log files in the directory ./logs