AbhishekRS4 / DRL_assignments

Deep Reinforcement Learning

Deep Reinforcement Learning assignments

Info

• Code for the assignments for the Deep Reinforcement Learning course offered at University of Groningen

Instructions to run code

Value based DRL

• Deep Q-Network (DQN) has been implemented from scratch
• Some code is inspired from this repo
• Run value_based_drl/train_dqn_agent.py for training
• Run value_based_drl/plot_success_rates.py to plot the success rate graph which is nothing but the average of the rewards of k test episodes

Policy based DRL

• Proximal Policy Optimization (PPO) has been implemented using StableBaselines3
• Run policy_based_drl/train_ppo_agent.py for training
• Run policy_based_drl/test_ppo_agent.py for testing
• Download csv file from the tensorboard logs and then run policy_based_drl/plot_graph.py

Package dependencies

• Can be found in requirements.txt