-
Notifications
You must be signed in to change notification settings - Fork 0
/
train_1.py
131 lines (118 loc) · 5.41 KB
/
train_1.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
131
import torch
import torch.nn as nn
from torchvision import datasets,models,transforms
import os
import time
import copy
import torch.optim as optim
if __name__ == '__main__':
img_size = 224
data_path = '..\\harvard_data\\train_val_split_data'
device = torch.device('cuda:0' if torch.cuda.is_available() else 'cpu')
print(device)
data_aug = {
'train': transforms.Compose([
transforms.RandomResizedCrop(img_size),
transforms.RandomHorizontalFlip(),
transforms.RandomVerticalFlip(),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
# transforms.Lambda(lambda crops: torch.stack([ToTensor()(crop) for crop in crops])),
]),
'val': transforms.Compose([
transforms.RandomResizedCrop(img_size),
transforms.RandomHorizontalFlip(),
transforms.RandomVerticalFlip(),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
# transforms.Lambda(lambda crops: torch.stack([ToTensor()(crop) for crop in crops])),
]),
}
# The images are stored in a dedicated folder for each class in data_path/train and data_path/val
# images['train'] contains tuples of image and label automatically given (scalar) for each class
# where the images for class 'ant' are stored in data_path/train/ant
images = {x:datasets.ImageFolder(os.path.join(data_path,x),data_aug[x])
for x in ['train','val']}
# DataLoader divides these into shuffled batches
dataloaders = {x:torch.utils.data.DataLoader(images[x],batch_size=4,shuffle=True,num_workers=4)
for x in ['train','val']}
print(images['train'].classes)
data_sizes = {x:len(images[x])
for x in ['train','val']}
def train_model(model,criterion,optimiser,scheduler,epochs):
best_val_model = copy.deepcopy(model.state_dict())
best_acc = 0.0
stats = {
'train': {
'loss':[],
'acc':[]
},
'val': {
'loss':[],
'acc':[]
}
}
since = time.time()
for epoch in range(epochs):
print(f'Epoch: {epoch+1}/{epochs}')
print('-'*10)
for phase in ['train','val']:
if phase=='train':
model.train()
else :
model.eval()
running_loss = 0.0
running_strikes = 0
for _X,_y in dataloaders[phase]:
_X,_y = _X.to(device),_y.to(device)
optimiser.zero_grad()
with torch.set_grad_enabled(phase=='train'):
# bs, ncrops, c, h, w = _X.size()
outputs = model(_X)
# outputs = outputs.view(bs,ncrops,-1).mean(1)
preds = torch.argmax(outputs,dim=1)
_ground_truths = _y.data
loss = criterion(outputs,_ground_truths)
running_loss += loss.item()*(_X.shape[0])
running_strikes += torch.sum(preds==_ground_truths)
if phase == 'train':
loss.backward()
optimiser.step()
epoch_loss = running_loss/data_sizes[phase]
epoch_acc = running_strikes*1.0/data_sizes[phase]
stats[phase]['loss'].append(epoch_loss)
stats[phase]['acc'].append(epoch_acc)
if phase == 'val':
scheduler.step(epoch_loss)
if epoch_acc > best_acc:
best_acc = stats[phase]['acc'][-1]
best_val_model = copy.deepcopy(model.state_dict())
print('{} Loss: {:.4f} Acc: {:.4f}'.format(
phase, epoch_loss, epoch_acc))
print('-'*40)
time_elapsed = time.time() - since
print('Training complete in {:.0f}m {:.0f}s'.format(
time_elapsed // 60, time_elapsed % 60))
print('Best val Acc: {:4f}'.format(best_acc))
model.load_state_dict(best_val_model)
return model,stats
model = models.resnet18(pretrained=True)
num_ftrs = model.fc.in_features
for param in model.parameters():
param.requires_grad = False
model.fc = nn.Sequential(nn.Linear(num_ftrs,256),nn.ReLU(),nn.Dropout(0.4),nn.Linear(256,4))
model.load_state_dict(torch.load('logs\\train_2_resnet18_lr=0.001_frozen.pt'))
model = model.to(device)
criterion = nn.CrossEntropyLoss()
optimiser = optim.SGD(model.parameters(),lr=0.001,momentum=0.9,nesterov=True)
scheduler = optim.lr_scheduler.ReduceLROnPlateau(optimiser,patience=4,mode='min',factor=0.2)
model, stats = train_model(model,criterion,optimiser,scheduler,40)
f = open('logs\\train_2_resnet18_lr=0.001_frozen.log','w')
for i in range(40):
tl = stats['train']['loss'][i]
ta = stats['train']['acc'][i]
vl = stats['val']['loss'][i]
va = stats['val']['acc'][i]
f.write(f'{i},{tl},{ta},{vl},{va}\n')
f.close()
torch.save(model.state_dict(), 'logs\\train_2_resnet18_lr=0.001_frozen.pt')