Data generation

Data generation

1. clone the repository
2. make

if make fails try:

make build=release

after you can give multiple parameters to data generation

./smallbrain -gen -threads <n> -book <path/to/book> -tb <path/to/tb> -depth <d>

there will be n amount of .txt's in ./data.

The format is

<fen> [<WDL>] <eval>

WDL
1 if White wins
0.5 for draws
0 for Black wins

Eval
Eval is always from Whites perspective.