Skip to content
This repository has been archived by the owner on Mar 2, 2022. It is now read-only.

Latest commit

 

History

History
403 lines (305 loc) · 17.8 KB

README.md

File metadata and controls

403 lines (305 loc) · 17.8 KB

CircleCI

Polygames

This README is a work in progress, please feel very free to post issues - we are happy to help. Save up computational power: you can find checkpoints here: http://dl.fbaipublicfiles.com/polygames/checkpoints/list.txt (feel free to open an issue for discussing which checkpoint you should use for which game/problem!).

For Nix users: see this doc.

Requirement:

C++17 compatible compiler
miniconda3

Compilation Guide:

First install conda and pytorch

Create a fresh conda environment with python3.7, install pytorch and dependencies.

# create a fresh conda environment with python3
# you will need to have miniconda3 set up
conda create --name [your env name] python=3.7 pip

conda activate [your env name] # Or source activate [your env name], depending on conda version.

conda install numpy pyyaml mkl mkl-include setuptools cmake cffi typing
conda install pytorch cudatoolkit=10.1 -c pytorch
conda install -c conda-forge tensorboardx
conda install -c conda-forge openjdk  # optional
conda install -c conda-forge graphviz # optional

pip install visdom
pip install torchviz				  # optional

Clone the repo and build

git clone --recursive https://github.com/facebookincubator/polygames
cd polygames

mkdir build
cd build

cmake .. -DCMAKE_BUILD_TYPE=relwithdebinfo -DPYTORCH15=ON
make -j

Ludii support can be disabled by appending -DWITH_LUDII=OFF to the cmake command (required if you don't have jdk)

Content

The repo contains mostly the following folders:

  • the pypolygames python package, which serves as an entry point for the application
  • the src folder, containing all C++ source code and third party libraries
    • the src/games folder, containing the games coded in C++

How to use the application

The application is launched from the pypolygames python package, in either of the following modes:ar

  • pypolygames train (training mode): a game and a model (as well as several other options, see below) are chosen and the model is iteratively trained with MCTS
  • pypolygames eval (evaluation mode): the model confronts either a pure MCTS or another neural network powered MCTS. The evaluation of a training can be done either offline (from checkpoints periodically saved) or in real time; in that case, the evaluation considers only the most recent checkpoint in order to follow closely the training, skipping some checkpoints in case the eval computation takes longer than the time becween consecutive checkpoints. It is displayed through visdom.
  • pypolygames traineval (training + evaluation mode): it mixes the two previous modes and allow to launch one command instead of two. With the real_time option the modes can be launched in parallel instead of sequentially.
  • pypolygames human (human mode): a human player plays against the machine

When a training is launched, it creates a game_GAMENAME_model_MODELNAME_feat_FEATURIZATION_GMT_YYYYMMDDHHMMSS within the save_dir where it will log relevant files:

  • model.pt
  • train.log
  • stat.tb
  • checkpoints_EPOCH.pt for for checkpoints saved each saving_period epoch (e.g., if saving_period == 10, checkpoints_0.pt, checkpoints_10.pt, checkpoints_20.pt, checkpoints_30.pt)

This directory will be the checkpoint_save_dir directory used by evaluation to retrieve the checkpoints to perform eval computation.

Parameters

The list of parameters for each mode is available with

python -m pypolygames {train,eval,traineval,human} --help

Threads

In train (resp. eval) mode, num_game * num_actor (resp. num_game * num_actor_eval * num_actor_opponent) is the total number of threads. The more num_actor (and num_actor_eval, num_actor_opponent), the larger the MCTS is for a given player.

In human mode, since num_game is set to one, for leveraging the computing power available on the platform, a rule-of-thumb is to set num_actor to 5 times the number of CPUs available (it is platform-dependent though, and performance tests should be done).

Model zoo

All models can be found in pypolygames/model_zoo. They come with a set of sensible parameters that can be customized as well as default games.

Usually models come in pair: MODELNAMEFCLogitModel and MODELNAMEConvLogitModel:

  • FCLogit models use a fully-connected layer for logit inference and are compatible with all games
  • ConvLogit models use a convolutional layer for logit inference and are only compatible with games whose action space if of same dimensions than their input space (an exception will be raised in case of an attempt to use an incompatible game)

So far the models being implemented are the folling:

  • GenericModel: generic model compatible with all games, default when no model_name is specified
  • NanoFCLogitModel: a simple model with a logit-inference fully-connected layer
  • NanoConvLogitModel: a simple model with a logit-inference convolutional layer
  • ResConvFCLogitModel: resnets with a logit-inference fully-connected layer
  • ResConvConvLogitModel: resnets with a logit-inference convolutional layer
  • UConvFCLogitModel: unets (direct paths between first and last layers) with a logit-inference fully-connected layer
  • UConvConvLogitModel: unets (direct paths between first and last layers) with a logit-inference convolutional layer
  • AmazonsModel: only for the Amazons game

Depending on the actual model chosen, some parameters might not have any use.

Featurization

--out_features=True: the input to the NN includes a channel with 1 on the frontier.
--turn_features=True: the input to the NN includes a channel with the player index broadcasted.
--geometric_features=True: the input to the NN includes 4 geometric channels representing the position on the board.
--random_features=4: the input to the NN includes 4 random features.
--one_feature=True: the input to the NN includes a channel with 1 everywhere.
--history=3: the representation from the last 3 steps is added in the featurization.

Examples

Run the following command before running the code

export OMP_NUM_THREADS=1

Examples for the training mode

  • Launch the game Connect4 with the GenericModel
python -m pypolygames train --game_name="Connect4"
  • Launch a game with a specific model and specific parameters
python -m pypolygames train --game_name="Connect4" --out_features=True \
    --model_name="UConvFCLogitModel" \
    --nnsize=16 \
    --nnks=3 \
    --pooling
  • Save checkpoints every 20 epochs in a specific folder
python -m pypolygames train --game_name="Connect4" --model_name="UConvFCLogitModel" \
    --saving_period=20 \
    --save_dir="/checkpoints"
  • Run training on GPU for a max time
python -m pypolygames train --game_name="Connect4" --model_name="UConvFCLogitModel" \
    --device="cuda:0" \
    --max_time=3600
  • Resume training from a given epoch
python -m pypolygames train \
    --save_dir="/checkpoints/game_Connect4_model_GenericModel_feat..._GMT_20190717103728" \
    --init_epoch=42
  • Initiate from a pretrained model
python -m pypolygames train --init_checkpoint="path/to/pretrained_model.pt" \
    --lr=0.001

Note that any checkpoint can serve as a pretrained model

  • Train on multiple GPUs
python -m pypolygames train --init_checkpoint "path/to/pretrained_model.pt" \
    --device cuda:0 cuda:1 cuda:2 cuda:3 cuda:4

In this case cuda:0 will be used for training the model while cuda:1, cuda:2 and cuda:3 will be used for generating games. If there is only one device specified, it will be used for both purposes.

Notes:

  • By default, the number of threads used for processing and batch sizes for inference are set automatically. These can be overriden with num_thread and per_thread_batchsize respectively.
  • num_game specifies the number of "master" threads scheduling games, and the total number of games being run in parallel will be num_game * per_thread_batchsize. Since per_thread_batchsize is automatically determined by default, this could be a large number in some instances.

Examples for the evaluation mode

  • Run offline evaluation
python -m pypolygames eval \
    --checkpoint_dir="/checkpoints/game_Connect4_model_GenericModel_feat..._GMT_20190717103728"
  • Plot evaluation on http://localhost:10000 as the same time as training happens (training needs to be run from another process)
python -m pypolygames eval \
    --checkpoint_dir="/checkpoints/game_Connect4_model_GenericModel_feat..._GMT_20190717103728" \
    --real_time \
    --plot_enabled \
    --plot_port=10000
  • Run evaluation on cpu with 100 games per evaluation, the pure-MCTS opponent playing 1000 rollouts while the model plays 400 rollouts
python -m pypolygames eval \
    --checkpoint_dir="/checkpoints/game_Connect4_model_GenericModel_feat..._GMT_20190717103728" \
    --device_eval="cpu" \
    --num_game_eval=100 \
    --num_rollouts_eval=400 \
    --num_actor_eval=8 \
    --num_rollouts_opponent=1000 \
    --num_actor_opponent=8
  • A specific checkpoint plays against another neural-network-powered MCTS
python -m pypolygames eval \
    --checkpoint="/checkpoints/checkpoint_600.zip" \
    --num_rollouts_eval=400 \
    --num_actor_eval=8 \
    --checkpoint_opponent="/checkpoints/checkpoint_200.zip" \
    --num_rollouts_opponent=1000 \
    --num_actor_opponent=8
  • Four GPUs are used for evaluating the model, all for inference
python -m pypolygames eval \
    --checkpoint="/checkpoints/checkpoint_600.zip" \
    --device_eval cuda:0 cuda:1 cuda:2 cuda:3 \
    --num_rollouts_eval=400 \
    --num_actor_eval=8 \
    --num_rollouts_opponent=1000 \
    --num_actor_opponent=8

Notes:

  • num_actor_eval, num_rollouts_eval, num_actor_opponent and num_rollouts_opponent are independent from the values used during training; in particular for proper benchmarking num_actor_eval and num_rollouts_eval should be set to the values used in human mode
  • num_game_eval * num_actor_eval (resp. num_game_eval * num_actor_opponent) is the number of threads used by the model to be evaluated (resp. the opponent)
  • there is no per_thread_batchsize in this mode
  • the higher num_actor_eval (resp. num_actor_opponent), the larger MCTS for a move in a given game will be, up to a limit where overheads between threads lead to decreasing returns. Empiracally this limit seems to be around 8. This limit may be game/model/platform dependent and should be tuned for a given instance.
  • against a pure MCTS opponent, num_rollouts_opponent should be set significantly higher than num_rollouts_eval

Examples for the training+evaluation mode

  • Run first training then evaluation on the last checkpoint
python -m pypolygames traineval --game_name="Connect4" \
    --save_dir="/checkpoints" \
    --num_epoch=1000
  • Plot evaluation on http://localhost:10000 as the same time as training happens
python -m pypolygames traineval --game_name="Connect4" \
    --save_dir="/checkpoints" \
    --real_time \
    --plot_enabled \
    --plot_port=10000

Examples for the human mode

  • Play to Connect4 against a pure MCTS as the second player with 8 threads
python -m pypolygames human --game_name="Connect4" \
    --pure_mcts \
    --num_actor 8
  • Play to Connect4 against a pretrained model as the second player
python -m pypolygames human \
    --init_checkpoint="/checkpoints/checkpoint_600.zip" \
    --human_first
  • Play with a timer, each side having 1800s in total, and the model playing each move with 0.07 of the remaining time
python -m pypolygames human \
    --init_checkpoint="/checkpoints/checkpoint_600.zip" \
    --total_time=1800 \
    --time_ratio=0.07
  • The model uses four GPUs, all for inference
python -m pypolygames human \
    --init_checkpoint "/checkpoints/checkpoint_600.zip" \
    --device cuda:0 cuda:1 cuda:2 cuda:3
  • The model uses four GPUs, all for inference, and uses the text protocol (actions are represented by x y z, each on one line):
python -m pypolygames tp \
    --init_checkpoint "/checkpoints/checkpoint_600.zip" \
    --device cuda:0 cuda:1 cuda:2 cuda:3

Notes:

  • in human mode, the model being fixed, the goal is to maximize performance given the platform running the model
  • the most effective way to improve model performance is to increase the MCTS size
  • as for training and evaluation, but given that there is only one game played, num_actor is the total number of threads
  • the higher num_actor, the larger the MCTS, up to a limit where overheads between threads lead to decreasing returns. Empiracally this limit seems to be around 8. This limit may be game/model/platform dependent and should be tuned for a given instance.
  • in a time-limited game num_rollouts should not be specified as it is maximized within each time_ratio * remaining time period

Examples for converting models

Saved checkpoints of models also store details about the game for which they were trained, and can only be used directly for the game in which they were trained. This is why eval runs do not require the --game_name to be specified; this is inferred from the model. The pypolygames convert command can be used to convert models to different games.

  • Fully automated convert between games:
python -m pypolygames convert \
    --init_checkpoint "/checkpoints/checkpoint_600.pt.gz" \
	--game_name="LudiiGomoku.lud" \
	--out="/checkpoints/converted/XToGomoku.pt.gz"

This takes the previously-trained model stored in "/checkpoints/checkpoint_600.pt.gz", modifies it such that it can be used to play the Ludii implementation of Gomoku, and stores this modified version of the model in the new file "/checkpoints/converted/XToGomoku.pt.gz".

This works best when using neural network architectures that are compatible with arbitrary board shapes (such as ResConvConvLogitPoolModel), and source and target games that have identical numbers of channels for state and move tensors, as well as identical semantics for those channels. For instance, the Ludii implementation of Yavalath has the same number of channels with identical semantics (in the same order) as Gomoku. Therefore, if the source model in "/checkpoints/checkpoint_600.pt.gz" was trained using --model_name=ResConvConvLogitPoolModel and --game_name="LudiiYavalath.lud", this conversion can be performed directly without having to delete any parameters or add any new parameters.

  • Fully automated convert between game options:
python -m pypolygames convert \
    --init_checkpoint "/checkpoints/checkpoint_600.pt.gz" \
	--game_options="Board Size/19x19" \
	--out="/checkpoints/converted/Gomoku/15x15_to_19x19.pt.gz"

This example will convert the source checkpoint "/checkpoints/checkpoint_600.pt.gz" into a model that can be used in a game loaded with the additional --game_options="Board Size/19x19" argument. For example, --game_name=LudiiGomoku.lud is by default played on a 15x15 board, but can be played on a larger 19x19 board with the --game_options="Board Size/19x19" argument.

Note that the convert command only takes game options into account if some form of --game_options is explicitly provided among the command line arguments. This means that, if a model was first trained for --game_options=Board Size/19x19, and the goal is to convert it into one for the default board size of 15x15, it is still necessary to provide either --game_options (without any values after it) or --game_options=Board Size/15x15 to the convert script. This tells it that the goal is indeed to revert to default options, rather than just leaving whichever options were baked into the source model.

Examples for generating figures of models

If the optional graphviz and torchviz dependencies are installed, we can use torchviz to automatically generate figures of our models. This can be done using draw_model script:

python -m pypolygames draw_model \
	--game_name="Hex5pie" \
	--model_name="ResConvConvLogitPoolModelV2" \
	--out="/private/home/$USER/ImageName"

This command will generate an image of the ResConvConvLogitPoolModelV2 architecture when playing Hex5pie, and save it to /private/home/$USER/ImageName.png (note that the .png extension will be automatically appended).

Any arguments that can be used to modify the game, or any aspect of the Neural Network architecture, can be used in this command.

Running games through Ludii

See detailed documentation on the Ludii integration here.

Contributing

We welcome contributions! Please check basic instructions here

Initial contributors

Contributors to the early version of Polygames (before open source release) include:

Tristan Cazenave, Univ. Dauphine; Yen-Chi Chen, National Taiwan Normal University; Guan-Wei Chen, National Dong Hwa University; Shi-Yu Chen, National Dong Hwa University; Xian-Dong Chiu, National Dong Hwa University; Julien Dehos, Univ. Littoral Cote d’Opale; Maria Elsa, National Dong Hwa University; Qucheng Gong, Facebook AI Research; Hengyuan Hu, Facebook AI Research; Vasil Khalidov, Facebook AI Research; Chen-Ling Li, National Dong Hwa University; Hsin-I Lin, National Dong Hwa University; Yu-Jin Lin, National Dong Hwa University; Xavier Martinet, Facebook AI Research; Vegard Mella, Facebook AI Research; Jeremy Rapin, Facebook AI Research; Baptiste Roziere, Facebook AI Research; Gabriel Synnaeve, Facebook AI Research; Fabien Teytaud, Univ. Littoral Cote d’Opale; Olivier Teytaud, Facebook AI Research; Shi-Cheng Ye, National Dong Hwa University; Yi-Jun Ye, National Dong Hwa University; Shi-Jim Yen, National Dong Hwa University; Sergey Zagoruyko, Facebook AI Research

License

polygames is released under the MIT license. See LICENSE for additional details about it. Third-party libraries are also included under their own license.