Restricted Boltzmann Machine

This repository is an implementation of Restricted Boltzmann Machine in Pytorch and Tensorflow.

RBM Structure

Getting started

Requirements

python 3.9
Tensorflow == 2.11.0
torch == 1.13.1
pandas == 1.5.3
numpy == 1.24.2
mlxtend == 0.21.0
sklearn == 1.1.1
matplotlib == 3.6.3

Usage

from RBM import RBM

rbm = RBM(nv=len(training_set[0]), nh=64, k=5, lr=0.01)
rbm.fit(training_set, batch_size=128, epoch=40, verbose=1, validation_data=validation_set)
predicted = rbm.predict(test_set, verbose=1)

RunTime reports

Visualize loss curves

import pandas as pd
pd.DataFrame(rbm.total_losses).plot()

Loss curves for minist dataset reconstruction using RBM

Visualize reconstructed images

import matplotlib.pyplot as plt

plt.figure(figsize=(12, 12))
for i in range(1, 33):
    plt.subplot(8, 8, i * 2 - 1)
    plt.imshow(predicted[i].reshape(28, 28))
    plt.subplot(8, 8, i * 2)
    plt.imshow(test_set[i].reshape(28, 28))

plt.show()

Reconstruction example of minist dataset

Visualize weights

import matplotlib.pyplot as plt

# visualize tensorflow weights
plt.figure(figsize=(12, 12))
for i in range(1, 65):
    plt.subplot(8, 8, i)
    plt.imshow(rbm.W[i-1].numpy().reshape(28, 28))
    
# visualize torch weights
plt.figure(figsize=(12, 12))
for i in range(1, 65):
    plt.subplot(8, 8, i)
    plt.imshow(rbm.W[i-1].reshape(28, 28))

plt.show()

Example weights for minist dataset

License

This repository is released under Apache License V2. To develop, publication and use it, please follow the terms of this license.

Citation

Please cite this paper if you use this code in your research project.

@article{
  title={An introduction to restricted Boltzmann machines.},
  author={A Fischer, C Igel},
  journal={Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 17th Iberoamerican Congress, CIARP 2012, Buenos Aires, Argentina, September 3-6, 2012. Proceedings 17},
  pages={14-36},
  year={2012},
  publisher={Springer}
}

arashazimi0032/Restricted-Boltzmann-Machine