if len(sys.argv) < 2:
# Choose the best one by the lowest training reconstruction error in last 10 epochs
# filename = 'MNIST_LeNet5DoubleFlatten_HalfCauchy-LogNormal-tau_conv_global(1.00E-04)tau_conv_local(3.00E-03)tau_fc_global(1.00E-04)tau_fc_local(3.00E-03)_E0200_01:20:10:600426_e0197.pkl'
# filename = 'MNIST_LeNet5DoubleFlatten_HalfCauchy-LogNormal-tau_conv_global(1.00E-04)tau_conv_local(1.00E-03)tau_fc_global(1.00E-04)tau_fc_local(1.00E-03)_E0200_01:20:13:088725_e0196.pkl'
# filename = 'MNIST_LeNet5DoubleFlatten_HalfCauchy-LogNormal-tau_conv_global(1.00E-04)tau_conv_local(5.00E-04)tau_fc_global(1.00E-04)tau_fc_local(5.00E-04)_E0200_01:20:18:234706_e0191.pkl'
# filename = 'MNIST_LeNet5DoubleFlatten_HalfCauchy-LogNormal-tau_conv_global(1.00E-04)tau_conv_local(1.00E-04)tau_fc_global(1.00E-04)tau_fc_local(1.00E-04)_E0200_01:20:21:680255_e0194.pkl'
# filename = 'MNIST_LeNet5DoubleFlatten_HalfCauchy-LogNormal-tau_conv_global(1.00E-04)tau_conv_local(5.00E-03)tau_fc_global(1.00E-04)tau_fc_local(5.00E-03)_E0200_01:20:27:171604_e0199.pkl'
# filename = 'MNIST_LeNet5DoubleFlatten_HalfCauchy-LogNormal-tau_conv_global(1.00E-04)tau_conv_local(1.00E-02)tau_fc_global(1.00E-04)tau_fc_local(1.00E-02)_E0200_12:08:27:167291_e0197.pkl'
# filename = 'MNIST_LeNet5DoubleFlatten_HalfCauchy-LogNormal-tau_conv_global(1.00E-04)tau_conv_local(5.00E-02)tau_fc_global(1.00E-04)tau_fc_local(5.00E-02)_E0200_12:08:12:007496_e0191.pkl'
# filename = 'MNIST_LeNet5DoubleFlatten_HalfCauchy-LogNormal-tau_conv_global(1.00E-04)tau_conv_local(1.00E-01)tau_fc_global(1.00E-04)tau_fc_local(1.00E-01)_E0200_16:30:05:309278_e0193.pkl'
filename = 'MNIST_LeNet5DoubleFlatten_E0200_e0002.pkl'
else:
filename = sys.argv[1]
row_threshold = [2.0, -8.0, -10.0, -9.0]
col_threshold = [1.0, -8.0, -11.0, -9.6]
model_file = os.path.join(exp_dir(), os.path.split(filename)[1])
dirname, filename = os.path.split(model_file)
model = load_model(model_type=model_type,
prior_info=prior_info_from_json('HalfCauchy.json'),
use_gpu=False)
model.load_state_dict(torch.load(model_file))
train_loader, valid_loader, test_loader, train_loader_eval = load_data(
data_type=data_type, batch_size=100, num_workers=0, use_gpu=False)
evaluate_with_prunning(model,
train_loader_eval,
valid_loader,
test_loader,
row_threshold=row_threshold,
col_threshold=col_threshold,
tag=filename)
import compress import numpy as np # Real training train = 'Data/Train/' small = 'Data/small/' X = compress.load_data(small) compress.compress_images(X, 100)
import numpy as np import pca as p import compress as c TRAINING_DATA = "Data/Train/" TEST_DATA = "Data/Test/" X = c.load_data(TRAINING_DATA) # c.compress_images(X, 10) c.compress_images(X, 100) # c.compress_images(X, 500) # c.compress_images(X, 1000) # c.compress_images(X, 2000) # c.compress_images(X, 50) # X = c.load_data(TEST_DATA) # c.compress_images(X, 10) # c.compress_images(X, 100) # c.compress_images(X, 500) # c.compress_images(X, 1000) # c.compress_images(X, 2000) # X = np.array([[-1, -1], [-1, 1], [1, -1], [1, 1]]) # X = np.array([[1, 1], [2, 7], [3, 3], [4, 4], [5, 5]]) # Z = p.compute_Z(X) # COV = p.compute_covariance_matrix(Z) # L, PCS = p.find_pcs(COV) # Z_star = p.project_data(Z, PCS, L, 2, 0) # print(Z_star) exit()
import pca
import numpy as np
import compress
# X = np.array([[1, 1], [1, 0], [2, 2], [2, 1], [2, 4], [3, 4], [
# 3, 3], [3, 2], [4, 4], [4, 5], [5, 5], [5, 7], [5, 4]])
# Z = pca.compute_Z(X, True, True)
# COV = pca.compute_covariance_matrix(Z)
# # print(COV)
# L, PCS = pca.find_pcs(COV)
# Zstar = pca.project_data(Z, PCS, L, 1, 0)
# print(PCS)
# print(Zstar)
X = compress.load_data('Data/Train/')
compress.compress_images(X, 10)
compress.compress_images(X, 100)
compress.compress_images(X, 500)
compress.compress_images(X, 1000)
compress.compress_images(X, 2000)
###################################################################################################################################################
# filename: PCATest.py
# author: Sara Davis
# date: 12/3/2018
# version: 1.0
# description: Run PCA.py
###########################################################################################################################
import pca
import numpy as np
import compress
X = np.array([[1, 1], [1,-1], [-1, 1], [-1, -1]])
centering = True
scaling = False
Z = pca.compute_Z(X, centering, scaling)
COV = pca.compute_covariance_matrix(Z)
L, PCS = pca.find_pcs(COV)
Z_star = pca.project_data(Z, PCS, L, 1, 0)
X = compress.load_data('/home/sara/Desktop/Data/Train/')
compress.compress_images(X, 100)
# -*- coding: utf-8 -*-
"""
Created on Thu Nov 21 15:03:00 2019
@author: pom_p
"""
import compress as comp
import numpy as np
X = comp.load_data('Train')
comp.compress_images(X, 1000)