Setup

Install PyTorch

python3 -m venv env
source env/bin/activate
pip install python-chess==0.31.4 pytorch-lightning==1.9.5 torch matplotlib tensorboard

Install CuPy

First check what version of cuda is being used by pytorch.

import torch
torch.version.cuda

Then install CuPy with the matching CUDA version.

pip install cupy-cudaXXX

where XXX corresponds to the first 3 digits of the CUDA version. For example cupy-cuda112 for CUDA 11.2.

CuPy might use the PyTorch's private installation of CUDA, but it is better to install the matching version of CUDA separately. CUDA Downloads

Build the fast DataLoader

This requires a C++17 compiler and cmake.

Windows:

compile_data_loader.bat

Linux/Mac:

sh compile_data_loader.bat

Train a network

source env/bin/activate
python train.py --gpus 1 train_data.bin val_data.bin

Resuming from a checkpoint

python train.py --gpus 1 --resume_from_checkpoint <path> ...

Feature set selection

By default the trainer uses a factorized HalfKAv2_hm feature set (named "HalfKAv2_hm^") If you wish to change the feature set used then you can use the --features=NAME option. For the list of available features see --help The default is:

python train.py ... --features="HalfKAv2_hm^"

Skipping certain fens in the training

--smart-fen-skipping currently skips over moves where the king is in check, or where the bestMove is a capture (typical of non-quiet positions). --random-fen-skipping N skip N fens on average before using one. Uses fewer fens per game, useful with large data sets.

Current recommended training invocation

python train.py --smart-fen-skipping --random-fen-skipping 3 --batch-size 16384 --threads 2 --num-workers 2 --gpus 1 trainingdata validationdata

best nets have been trained with 16B d9-scored nets, training runs >200 epochs

Export a network

Using either a checkpoint (.ckpt) or serialized model (.pt), you can export to SF NNUE format. This will convert last.ckpt to nn.nnue, which you can load directly in SF.

python serialize.py last.ckpt nn.nnue

Import a network

Import an existing SF NNUE network to the pytorch network format.

python serialize.py nn.nnue converted.pt

Visualize a network

Visualize a network from either a checkpoint (.ckpt), a serialized model (.pt) or a SF NNUE file (.nnue).

python visualize.py nn.nnue --features="HalfKAv2_hm"

Visualize the difference between two networks from either a checkpoint (.ckpt), a serialized model (.pt) or a SF NNUE file (.nnue).

python visualize.py nn.nnue  --features="HalfKAv2_hm" --ref-model nn.cpkt --ref-features="HalfKAv2_hm^"

Logging

pip install tensorboard
tensorboard --logdir=logs

Then, go to http://localhost:6006/

Automatically run matches to determine the best net generated by a (running) training

python run_games.py --concurrency 16 --stockfish_exe ./stockfish.master --c_chess_exe ./c-chess-cli --ordo_exe ./ordo --book_file_name ./noob_3moves.epd run96

Automatically converts all .ckpt found under run96 to .nnue and runs games to find the best net. Games are played using c-chess-cli and nets are ranked using ordo. This script runs in a loop, and will monitor the directory for new checkpoints. Can be run in parallel with the training, if idle cores are available.

Thanks

Sopel - for the amazing fast sparse data loader
connormcmonigle - https://github.com/connormcmonigle/seer-nnue, and loss function advice.
syzygy - http://www.talkchess.com/forum3/viewtopic.php?f=7&t=75506
https://github.com/DanielUranga/TensorFlowNNUE
https://hxim.github.io/Stockfish-Evaluation-Guide/
dkappe - Suggesting ranger (https://github.com/lessw2020/Ranger-Deep-Learning-Optimizer)

official-stockfish/nnue-pytorch

official-stockfish

Reviews

Repository Details