Is it yet another emulator? NO!

I wrote an AI-oriented, high-performance (5k FPS/i7 core, 500k FPS V100 GPU) and minimal Nintendo Gameboy emulator in C (~500 LOC) with 0 dependencies.

The goals of his project are:

1. provide an alternative to existing research datasets in Reinforcement Learning

2. provide much better performance (frames/second) than other emulators

3. give access to a platform with a variety of games (target meta- or transfer learning research)

4. enable modifications due to its simplicity

I provided an example python script communicating with the emulator, making usage almost 1:1 same as in case of OpenAI GYM.

There are over 1000 games only for Gameboy Classic (4 colors, 160x144), plus another 500 for the color version

FULL LIST OF GAMES (txt)

FULL LIST OF TESTED GAMES (GIFs)

Some notable examples, from left to right: Tetris, Micro Machines, Prince of Persia, Pacman, Space Invaders, Super Mario Land

Links

Slides PDF Slides Keynote

A GAMEBOY SUPERCOMPUTER: My medium article which started it all

Experiments with world models