Opponent Modelling library for StarCraft: Brood War, using BWAPI
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
OpponentModelling
PseudoBotForParsingReplays
.gitignore
README.md

README.md

bwom

Opponent Modelling library for StarCraft: Brood War, using BWAPI

Current state

The library is currently under developement. We are working on creating 9 independent neural networks for each matchup (PvP, PvZ, PvT, TvT, TvZ, TvP, ZvZ, ZvP, ZvT). Currently, there are implemented 3 neural networks for each race, but because of very high divergence in build orders and unit composition in every single matchup, 3 neural networks are not accurate enough.

Detailed description and usage with examples is provided in github wiki - https://github.com/Games-and-Simulations/bwom/wiki