This repo contains the implementation of the algorithm proposed in Off-Belief Learning, ICML 2021.

Overview

Off-Belief Learning

Introduction

This repo contains the implementation of the algorithm proposed in Off-Belief Learning, ICML 2021.

Environment Setup

We have been using pytorch-1.5.1, cuda-10.1, and cudnn-v7.6.5 in our development environment. Other settings may also work but we have not tested it extensively under different configurations. We also use conda/miniconda to manage environments.

There are known issues when using this repo with newer versions of pytorch, such as this illegal move issue.

conda create -n hanabi python=3.7
conda activate hanabi

# install pytorch 1.5.1
# note that newer versions may cause compilation issues
pip install torch==1.5.1+cu101 torchvision==0.6.1+cu101 -f https://download.pytorch.org/whl/torch_stable.html

# install other dependencies
pip install psutil

# install a newer cmake if the current version is < 3.15
conda install -c conda-forge cmake

To help cmake find the proper libraries (e.g. libtorch), please either add the following lines to your .bashrc, or add it to a separate file and source it before you start working on the project.

# activate the conda environment
conda activate hanabi

# set path
CONDA_PREFIX=${CONDA_PREFIX:-"$(dirname $(which conda))/../"}
export CPATH=${CONDA_PREFIX}/include:${CPATH}
export LIBRARY_PATH=${CONDA_PREFIX}/lib:${LIBRARY_PATH}
export LD_LIBRARY_PATH=${CONDA_PREFIX}/lib:${LD_LIBRARY_PATH}

# avoid tensor operation using all cpu cores
export OMP_NUM_THREADS=1

Finally, to compile this repo:

# under project root
mkdir build
cd build
cmake ..
make -j10

Code Structure

For an overview of how the training infrastructure, please refer to Figure 5 of the Off-Belief Learning paper.

hanabi-learning-environment is a modified version of the original HLE from Deepmind.

Notable modifications includes:

  1. Card knowledge part of the observation encoding is changed to v0-belief, i.e. card knowledge normalized by the remaining public card count.

  2. Functions to reset the game state with sampled hands.

rela (REinforcement Learning Assemly) is a set of tools for efficient batched neural network inference written in C++ with multi-threading.

rlcc implements the core of various algorithms. For example, the logic of fictitious transitions are implemented in r2d2_actor.cc. It also contains implementations of baselines such as other-play, VDN and IQL.

pyhanabi is the main entry point of the repo. It contains implementations for Q-network, recurrent DQN training, belief network and training, as well as some tools to analyze trained models.

Run the Code

Please refer to the README in pyhanabi for detailed instruction on how to train a model.

Download Models

To download the trained models used in the paper, go to models folder and run

sh download.sh

Due to agreement with BoardGameArena and Facebook policies, we are unable to release the "Clone Bot" models trained on the game data nor the datasets themselves.

Copyright

Copyright (c) Facebook, Inc. and its affiliates. All rights reserved.

This source code is licensed under the license found in the LICENSE file in the root directory of this source tree.

Owner
Facebook Research
Facebook Research
Signals-backend - A suite of card games written in Python

Card game A suite of card games written in the Python language. Features coming

1 Feb 15, 2022
Much faster than SORT(Simple Online and Realtime Tracking), a little worse than SORT

QSORT QSORT(Quick + Simple Online and Realtime Tracking) is a simple online and realtime tracking algorithm for 2D multiple object tracking in video s

Yonghye Kwon 8 Jul 27, 2022
Image Super-Resolution by Neural Texture Transfer

SRNTT: Image Super-Resolution by Neural Texture Transfer Tensorflow implementation of the paper Image Super-Resolution by Neural Texture Transfer acce

Zhifei Zhang 413 Nov 30, 2022
High performance distributed framework for training deep learning recommendation models based on PyTorch.

PERSIA (Parallel rEcommendation tRaining System with hybrId Acceleration) is developed by AI 340 Dec 30, 2022

Recovering Brain Structure Network Using Functional Connectivity

Recovering-Brain-Structure-Network-Using-Functional-Connectivity Framework: Papers: This repository provides a PyTorch implementation of the models ad

5 Nov 30, 2022
Tilted Empirical Risk Minimization (ICLR '21)

Tilted Empirical Risk Minimization This repository contains the implementation for the paper Tilted Empirical Risk Minimization ICLR 2021 Empirical ri

Tian Li 40 Nov 28, 2022
Christmas face app for Decathlon xmas coding party!

Christmas Face Application Use this library to create the perfect picture for your christmas cards! Done by Hasib Zunair, Guillaume Brassard and Samue

Hasib Zunair 4 Dec 20, 2021
A large-scale video dataset for the training and evaluation of 3D human pose estimation models

ASPset-510 (Australian Sports Pose Dataset) is a large-scale video dataset for the training and evaluation of 3D human pose estimation models. It contains 17 different amateur subjects performing 30

Aiden Nibali 25 Jun 20, 2021
Code of Adverse Weather Image Translation with Asymmetric and Uncertainty aware GAN

Adverse Weather Image Translation with Asymmetric and Uncertainty-aware GAN (AU-GAN) Official Tensorflow implementation of Adverse Weather Image Trans

Jeong-gi Kwak 36 Dec 26, 2022
Python based Advanced AI Assistant

Knick is a virtual artificial intelligence project, fully developed in python. The objective of this project is to develop a virtual assistant that can handle our minor, intermediate as well as heavy

19 Nov 15, 2022
CONditionals for Ordinal Regression and classification in tensorflow

Condor Ordinal regression in Tensorflow Keras Tensorflow Keras implementation of CONDOR Ordinal Regression (aka ordinal classification) by Garrett Jen

9 Jul 31, 2022
Monk is a low code Deep Learning tool and a unified wrapper for Computer Vision.

Monk - A computer vision toolkit for everyone Why use Monk Issue: Want to begin learning computer vision Solution: Start with Monk's hands-on study ro

Tessellate Imaging 507 Dec 04, 2022
Transfer Reinforcement Learning for Differing Action Spaces via Q-Network Representations

Transfer-Learning-in-Reinforcement-Learning Transfer Reinforcement Learning for Differing Action Spaces via Q-Network Representations Final Report Tra

Trung Hieu Tran 4 Oct 17, 2022
GAN-STEM-Conv2MultiSlice - Exploring Generative Adversarial Networks for Image-to-Image Translation in STEM Simulation

GAN-STEM-Conv2MultiSlice GAN method to help covert lower resolution STEM images generated by convolution methods to higher resolution STEM images gene

UW-Madison Computational Materials Group 2 Feb 10, 2021
[TPAMI 2021] iOD: Incremental Object Detection via Meta-Learning

Incremental Object Detection via Meta-Learning To appear in an upcoming issue of the IEEE Transactions on Pattern Analysis and Machine Intelligence (T

Joseph K J 66 Jan 04, 2023
Multiview Neural Surface Reconstruction by Disentangling Geometry and Appearance

Multiview Neural Surface Reconstruction by Disentangling Geometry and Appearance Project Page | Paper | Data This repository contains an implementatio

Lior Yariv 521 Dec 30, 2022
NanoDet-Plus⚡Super fast and lightweight anchor-free object detection model. 🔥Only 980 KB(int8) / 1.8MB (fp16) and run 97FPS on cellphone🔥

NanoDet-Plus⚡Super fast and lightweight anchor-free object detection model. 🔥Only 980 KB(int8) / 1.8MB (fp16) and run 97FPS on cellphone🔥

4.8k Jan 07, 2023
R-package accompanying the paper "Dynamic Factor Model for Functional Time Series: Identification, Estimation, and Prediction"

dffm The goal of dffm is to provide functionality to apply the methods developed in the paper “Dynamic Factor Model for Functional Time Series: Identi

Sven Otto 3 Dec 09, 2022
Official PyTorch implementation of "Meta-Learning with Task-Adaptive Loss Function for Few-Shot Learning" (ICCV2021 Oral)

MeTAL - Meta-Learning with Task-Adaptive Loss Function for Few-Shot Learning (ICCV2021 Oral) Sungyong Baik, Janghoon Choi, Heewon Kim, Dohee Cho, Jaes

Sungyong Baik 44 Dec 29, 2022
Aiming at the common training datsets split, spectrum preprocessing, wavelength select and calibration models algorithm involved in the spectral analysis process

Aiming at the common training datsets split, spectrum preprocessing, wavelength select and calibration models algorithm involved in the spectral analysis process, a complete algorithm library is esta

Fu Pengyou 50 Jan 07, 2023