bvk1ng's picture
Stage-1 commit: Agent trained for 3500 episodes
c121225
---
license: mit
language:
- en
tags:
- reinforcement learning
- games
---
# Deep Q-Learning based Agent for Atari Breakout
The agent showcased in this space is trained using the Deep Q-Learning algorithm.
The agent was trained for $3500$ episodes with a learning rate of $0.00001$ and an epsilon value that decreased linearly over time.