Skip to content

Latest commit

 

History

History
28 lines (19 loc) · 896 Bytes

README.md

File metadata and controls

28 lines (19 loc) · 896 Bytes

c4ai

A reinforcement learning artificial intelligence system used to play the game of connect-4.

How good it is

Good enough to smash all my friends, myself included pretty much all the time.

How bad it is

Bad enough to lose to a perfect player sometimes.

Known issues

  • schwi.py might not work on every system. This problem is a huge pain and I'm not interested in fixing it at the moment.
  • Specifying training arguments in do_selfplay of selfplay_v2.py doesn't change anything. I'm too lazy to do do that right now. Will definitely work on it later though...

By the way

  • training_pipeline.py will not use selfplay_v2.py by default. However, I have managed to (and so can you) build the binary on Linux. Use onnxruntime version 1.2.0 (see here) and recent version of GSL.
  • If you have something to add, cough it up.