vladfi1/phillip
The SSBM "Phillip" AI.
Trains deep reinforcement learning agents to play Super Smash Bros. Melee by interfacing directly with the Dolphin emulator, using custom communication protocols (ZMQ for training synchronization, TCP loopback on Windows) to enable real-time game state observation and action execution. Supports distributed training across SLURM clusters and includes pre-trained agents with varying playstyles, from frame-perfect to human-like behavior. Note: This project is unmaintained; see the successor slippi-ai for a modern imitation learning approach.
580 stars. No commits in the last 6 months.
Stars
580
Forks
79
Language
Python
License
GPL-3.0
Category
Last pushed
Jan 04, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/vladfi1/phillip"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
yenchenlin/DeepLearningFlappyBird
Flappy Bird hack using Deep Reinforcement Learning (Deep Q-learning).
ChenglongChen/pytorch-DRL
PyTorch implementations of various Deep Reinforcement Learning (DRL) algorithms for both single...
nikp06/subwAI
Scripts for training an AI to play the endless runner Subway Surfers using a supervised machine...
kingyuluk/RL-FlappyBird
Using reinforcement learning to train FlappyBird.
vita-epfl/social-nce
[ICCV] Social NCE: Contrastive Learning of Socially-aware Motion Representations