"""Training of Restricted Boltzmann Machine"""
model = RBM(n_visible=28 * 28,
n_hidden=144,
lr=0.1,
l2_coeff=1e-4,
momentum_coff=0.5)
training_epochs = 5
for epoch in range(training_epochs):
# index = 0
img_num = 60000
error = 0
for i in range(img_num):
x = x_train[i].reshape(1, 784)
error += model.const_divergence(x / 255.0, K=10)
error = error / img_num
print('Training epoch %d, cost is ' % epoch, error)
rbm_weight = open(b"model/rbm_weight.npy", "wb")
pickle.dump(model.weight, rbm_weight)
rbm_v_bias = open(b"model/rbm_v_bias.npy", "wb")
pickle.dump(model.v_bias, rbm_v_bias)
rbm_h_bias = open(b"model/rbm_h_bias.npy", "wb")
pickle.dump(model.h_bias, rbm_h_bias)
# """Plotting and Testing of Restricted Boltzmann Machine"""
#
import pickle
from struct import unpack
import gzip
import numpy as np
from RBM import RBM
#Dataset
f = gzip.open("../data/mnist.pkl.gz", "rb")
data = pickle.load(f, encoding='bytes')
f.close()
(x_train, y_train), (x_test, y_test) = data
# x_train = x_train.reshape(x_train.shape[0],x_train.shape[1]*x_train.shape[2])
x = x_train[0].reshape(1, 784)
"""Training"""
model = RBM(n_visible=28 * 28, n_hidden=200, lr=0.05)
training_epochs = 2
for epoch in range(training_epochs):
cost = model.const_divergence(x, K=1)
print('Training epoch %d, cost is ' % epoch, cost)
# end_time = time.clock()
# pretraining_time = (end_time - start_time)
# print ('Training took %f minutes' % (pretraining_time / 60.))
