Cory

All I Need is 60ms

Augmented RL

binbot
Alpha Beta Snake

🧬evolve 🧬

Fast Snake Clone

Heroku RL

🦡 honeybadger 🦡

Imperfect ML Snake
Reinforcement Learning Snake -- trained with custom gym environment & PPO with self play. Training setup modeled after OpenAI Dota 5

Imperfect ML Snake 2

Imperfect ML Snake 3

MCTS Experiment
Simple MCTS

ML Only (Reinforcement Learning) - 4 player tuned

♕ Serpent Queen ♕

Special Snake.
Specifically trained for 1v1. Pure ML, no tree searching. Currently running a dumb model.

Thinks fast or dies trying
Deep Neural Network Policy (PPO)