Skip to content

yifeng-li/xdgm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

5 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

This package includes Yifeng Li's implementations of exponential family deep generative models. New developments will be gradually added.

1. Implementations in Python 3.6:

restricted_boltzmann_machine.py: Class of exponential family restricted Boltzmann machine (exp-RBM).
helmholtz_machine.py: Class of exponential family Helmholtz machine (exp-HM).
deep_belief_net.py: Class of exponential family deep belief net (exp-DBN).
deep_boltzmann_machine.py: Class of exponential family deep Boltzman machine (exp-DBM).

multimodaldbm.py: Class of multi-modal deep Boltzmann machine (MDBM).
multimodaldbm.py: Class of multi-modal deep belief net (MDBN).


2. Examples:

main_test_ExpRBM_MNIST.py: Exp-RBM on MNIST.
main_test_ExpRBM_FASHIONMNIST.py: Exp-RBM on Fashion-MNIST.
main_test_ExpHM_MNIST.py: Exp-HM on MNIST.
main_test_ExpHM_FASHIONMNIST.py: Exp-HM on Fashion-MNIST.
main_test_ExpDBN_MNIST.py: Exp-DBN on MNIST.
main_test_ExpDBN_FASHIONMNIST.py: Exp-DBN on Fashion-MNIST.
main_test_ExpDBM_MNIST.py: Exp-DBM on MNIST.
main_test_ExpDBM_MNIST.py: Exp-DBM on MNIST.
main_test_ExpDBM_FASHIONMNIST.py: Exp-DBM on Fashion-MNIST.
main_test_ExpMDBN_MNIST.py: MDBN on MNIST.
main_test_ExpMDBN_FASHIONMNIST.py: MDBN on Fashion-MNIST.
main_test_ExpMDBM_MNIST.py: MDBM on MNIST.


3. Data:

MNIST: given in ./data/MNIST.
Fashion-MNIST: if missing, download the csv files from https://www.kaggle.com/zalando-research/fashionmnist, then added them to ./data/FASHIONMNIST.


4. Citations:

[1] Yifeng Li and Xiaodan Zhu, "Exploring Helmholtz machine and deep belief net in the exponential family perspective," ICML 2018 Workshop on Theoretical Foundations and Applications of Deep Generative Models, Stockholm, Sweden, July 2018. 

[2] Yifeng Li and Xiaodan Zhu, "Exponential family restricted Boltzmann machines and annealed importance sampling," 2018 International Joint Conference on Neural Networks (IJCNN/WCCI), Rio, Brazil, July 2018, pp. 39-48.


5. Contact:

Yifeng Li
Scientific Data Mining Team
Digital Technologies Research Centre
National Research Council Canada
Email: yifeng.li.cn@gmail.com; yifeng.li@nrc-cnrc.gc.ca
Web1: https://sites.google.com/view/yifengli
Web2: https://www.nrc-cnrc.gc.ca/eng/expertise/profile.html?id=33073


Other Packages:
Capsule Generative Models: https://github.com/yifeng-li/cdgm
Deep Learning: https://github.com/yifeng-li/DECRES
MVMF: https://github.com/yifeng-li/mvmf
NMF Toolbox: https://sites.google.com/site/nmftool
SR Toolbox:    https://sites.google.com/site/sparsereptool
RLMK Toolbox: https://sites.google.com/site/rlmktool
PGM Toolbox:   https://sites.google.com/site/pgmtool
Spectral Clustering Toolbox: https://sites.google.com/site/speclust

About

Exponential Family Deep Generative Models.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages