/
Neural_Network.py
130 lines (114 loc) · 4.67 KB
/
Neural_Network.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
import urllib
import warnings
import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from numpy.distutils.conv_template import header
from torchvision import datasets, transforms
def get_data_loader(training = True):
transform = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize((0.1307,), (0.3081,))
])
train_set = datasets.MNIST(root = 'data', train=True, download=True, transform=transform)
test_set = datasets.MNIST(root = 'data', train=False, download=False, transform=transform)
train_loader = torch.utils.data.DataLoader(train_set, batch_size=50)
test_loader = torch.utils.data.DataLoader(test_set, batch_size=50, shuffle=False)
if training:
return train_loader
else:
return test_loader
def build_model():
warnings.filterwarnings('ignore')
model = nn.Sequential(nn.Flatten(), nn.LazyLinear(128), nn.ReLU(), nn.LazyLinear(64), nn.ReLU(), nn.LazyLinear(10))
return model
def train_model(model, train_loader, criterion, T):
op = optim.SGD(model.parameters(), lr=0.001, momentum=0.9)
model.train()
for epoch in range(T):
running_loss = 0.0
total = 0
correct = 0
for data in train_loader:
inputs, labels = data
op.zero_grad()
outputs = model(inputs)
_, predicted = torch.max(outputs.data, 1)
total += labels.size(0)
correct += (predicted == labels).sum().item()
loss = criterion(outputs, labels)
loss.backward()
op.step()
running_loss += loss.item()
print("Train Epoch: %d Accuracy: %d/%d(%.2f%%) Loss: %.3f" %(epoch, correct, total, 100 * (correct / total), running_loss / len(train_loader)))
def evaluate_model(model, test_loader, criterion, show_loss=True):
model.eval()
running_loss = 0.0
total = 0
correct = 0
with torch.no_grad():
for data in test_loader:
inputs, labels = data
outputs = model(inputs)
_, predicted = torch.max(outputs.data, 1)
total += labels.size(0)
correct += (predicted == labels).sum().item()
loss = criterion(outputs, labels)
running_loss += loss.item()
if show_loss:
print("Average Loss: %.4f" %(running_loss / len(test_loader)))
print("Accuracy: %.2f%%" %(100 * (correct / total)))
else:
print("Accuracy: %.2f%%" % (100 * (correct / total)))
def predict_label(model, test_images, index):
final_prob = 0
class_names = ['zero', 'one', 'two', 'three', 'four', 'five', 'six', 'seven', 'eight', 'nine']
p_list = list()
p_dict = {}
output = model(test_images)
prob = F.softmax(output, dim=1)
img = list(prob[index])
predicted = sorted(img)[::-1]
for tensor in img:
p_list.append(tensor.item() * 100)
zero, one, two, three, four, five, six, seven, eight, nine = p_list[0], p_list[1], p_list[2], p_list[3], p_list[4], p_list[5], p_list[6], p_list[7], p_list[8], p_list[9]
for p in range(3):
if predicted[p].item() * 100 == zero:
print("zero: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == one:
print("one: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == two:
print("two: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == three:
print("three: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == four:
print("four: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == five:
print("five: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == six:
print("six: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == seven:
print("seven: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == eight:
print("eight: %.2f%%" %(predicted[p].item() * 100))
continue
if predicted[p].item() * 100 == nine:
print("nine: %.2f%%" %(predicted[p].item() * 100))
continue
if __name__ == '__main__':
criterion = nn.CrossEntropyLoss()
model = build_model()
train_model(model, get_data_loader(), criterion, T=5)
evaluate_model(model, get_data_loader(training=False), criterion, show_loss=True)
test_images, _ = iter(get_data_loader()).next()
predict_label(model, test_images, 1)