SAVER

This repository contains the code to run the experiment (plus some additional experiments) described in the workshop paper “Quantile Regression Reinforcement Learning with State Aligned Vector Rewards” - https://openreview.net/forum?id=S1fuWRYCFm

Requirements

• Python v3.5
• Tensorflow (https://www.tensorflow.org/install/)
• OpenAI Gym (https://github.com/openai/gym)

Installation

• Download the folder and install requirements
• Install and register the gym environments in the sub directory “saver_envs”
    ◦ $ pip install -e saver_envs
• Run the file “main.py” to train an agent
    ◦ $ python main.py