Skip to content

The code base for my Master thesis "Transformer based action sequence generation in reinforcement learning settings"

Notifications You must be signed in to change notification settings

Gideon-Stein/TBQN

Repository files navigation

TBQN

The code base for my Master thesis "Transformer based action sequence generation in reinforcement learning settings". It includes fully modularized code to train, Hp optimize and evaluate a DQN agent with a Transformer based architecture as its network (TBQN). The basic network structure is depicted below.

drawing

Getting Started

This repository includes the following useful things:

  • Fully modularized code to Run a DQN agent with a Transformer based architecture as its network (TBQN).
  • Simple scripts to run TBQN with a mountain of different Parameters and model variations.
  • Scripts to perform parameter optimization for TBQN using the Optuna Library
  • The code and the results of the experiments I conducted during my thesis work.
  • Notebooks that can be used to evaluate and display either single model performance or whole studies.

Build on

Installing

To install dependencies simply run

pip install -r requirements.txt

You should be good to go.

Example usage

To run an experiment or a study simply run one of the scripts like this:

python experiment_script_3.py --output_dir Acrobot-v1 --env Acrobot-v1   

Parameters can be added and changed accordingly to the script.

To evaluate the results of an experiment simply load your results into Result_display.ipynb. Inside the Notebook specify the saved experiment path:

directory = "experiment_scripts/example_result"

To evaluate the results of a study use Study_display.ipynb. or Study_display_2.ipynb.

Authors

  • Gideon Stein - Initial work - Github

Comments

  • I am currently working on an PPO version which is not finished yet. Therefore, PPO named files are not finalized neither complete. They should be ignored.
  • Due to storage I cannot include concrete studies and model runs in this repository. If you are interested in anything additional data, feel free to write me.

About

The code base for my Master thesis "Transformer based action sequence generation in reinforcement learning settings"

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published