Skip to content
A short and sweet implementation of an A2C agent to play Atari Breakout
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.

README.md

breakout-demo

Super simple, no-frills A2C agent that achieves over 200 reward on Atari Breakout, with MG2033/A2C and openai/baselines as reference.

Here's the corresponding blog post.

To train a model from scratch, run

python a2c.py

I recommend a value of N = 50 or 100 for best results, though training does take some time with those values.

python a2c.py --n 100

Better graphs, Tensorboard visualizations, testing, and saved model files on the way.

Results

N Max Reward Iterations before overfit
1
5
20 376
50 397 Less than 1020626
100 428 Less than 1031646
Inf
You can’t perform that action at this time.