-
Notifications
You must be signed in to change notification settings - Fork 178
/
train.py
157 lines (139 loc) · 6.42 KB
/
train.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
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
# -*- coding: utf-8 -*-
"""
@author: Viet Nguyen <[email protected]>
"""
import argparse
import os
import shutil
import numpy as np
import torch
import torch.nn as nn
from tensorboardX import SummaryWriter
from torch.utils.data import DataLoader
from src.dataset import MyDataset
from src.model import QuickDraw
from src.utils import get_evaluation
def get_args():
parser = argparse.ArgumentParser(
"""Implementation of the Quick Draw model proposed by Google""")
parser.add_argument("--optimizer", type=str, choices=["sgd", "adam"], default="sgd")
parser.add_argument("--total_images_per_class", type=int, default=10000)
parser.add_argument("--ratio", type=float, default=0.8, help="the ratio between training and test sets")
parser.add_argument("--batch_size", type=int, default=32)
parser.add_argument("--num_epochs", type=int, default=20)
parser.add_argument("--lr", type=float,
default=0.01) # recommended learning rate for sgd is 0.01, while for adam is 0.001
parser.add_argument("--es_min_delta", type=float, default=0.0,
help="Early stopping's parameter: minimum change loss to qualify as an improvement")
parser.add_argument("--es_patience", type=int, default=3,
help="Early stopping's parameter: number of epochs with no improvement after which training will be stopped. Set to 0 to disable this technique.")
parser.add_argument("--data_path", type=str, default="data", help="the root folder of dataset")
parser.add_argument("--log_path", type=str, default="tensorboard")
parser.add_argument("--saved_path", type=str, default="trained_models")
args = parser.parse_args()
return args
def train(opt):
if torch.cuda.is_available():
torch.cuda.manual_seed(123)
else:
torch.manual_seed(123)
training_params = {"batch_size": opt.batch_size,
"shuffle": True}
test_params = {"batch_size": opt.batch_size,
"shuffle": False}
output_file = open(opt.saved_path + os.sep + "logs.txt", "w")
output_file.write("Model's parameters: {}".format(vars(opt)))
training_set = MyDataset(opt.data_path, opt.total_images_per_class, opt.ratio, "train")
training_generator = DataLoader(training_set, **training_params)
print ("there are {} images for training phase".format(training_set.__len__()))
test_set = MyDataset(opt.data_path, opt.total_images_per_class, opt.ratio, "test")
test_generator = DataLoader(test_set, **test_params)
print("there are {} images for test phase".format(test_set.__len__()))
model = QuickDraw(num_classes=training_set.num_classes)
if os.path.isdir(opt.log_path):
shutil.rmtree(opt.log_path)
os.makedirs(opt.log_path)
writer = SummaryWriter(opt.log_path)
# writer.add_graph(model, torch.rand(opt.batch_size, 1, 28, 28))
if torch.cuda.is_available():
model.cuda()
criterion = nn.CrossEntropyLoss()
if opt.optimizer == "adam":
optimizer = torch.optim.Adam(model.parameters(), lr=opt.lr)
elif opt.optimizer == "sgd":
optimizer = torch.optim.SGD(model.parameters(), lr=opt.lr, momentum=0.9)
else:
print("invalid optimizer")
exit(0)
best_loss = 1e5
best_epoch = 0
model.train()
num_iter_per_epoch = len(training_generator)
for epoch in range(opt.num_epochs):
for iter, batch in enumerate(training_generator):
images, labels = batch
if torch.cuda.is_available():
images = images.cuda()
labels = labels.cuda()
optimizer.zero_grad()
predictions = model(images)
loss = criterion(predictions, labels)
loss.backward()
optimizer.step()
training_metrics = get_evaluation(labels.cpu().numpy(), predictions.cpu().detach().numpy(),
list_metrics=["accuracy"])
print("Epoch: {}/{}, Iteration: {}/{}, Lr: {}, Loss: {}, Accuracy: {}".format(
epoch + 1,
opt.num_epochs,
iter + 1,
num_iter_per_epoch,
optimizer.param_groups[0]['lr'],
loss, training_metrics["accuracy"]))
writer.add_scalar('Train/Loss', loss, epoch * num_iter_per_epoch + iter)
writer.add_scalar('Train/Accuracy', training_metrics["accuracy"], epoch * num_iter_per_epoch + iter)
model.eval()
loss_ls = []
te_label_ls = []
te_pred_ls = []
for idx, te_batch in enumerate(test_generator):
te_images, te_labels = te_batch
num_samples = te_labels.size()[0]
if torch.cuda.is_available():
te_images = te_images.cuda()
te_labels = te_labels.cuda()
with torch.no_grad():
te_predictions = model(te_images)
te_loss = criterion(te_predictions, te_labels)
loss_ls.append(te_loss * num_samples)
te_label_ls.extend(te_labels.clone().cpu())
te_pred_ls.append(te_predictions.clone().cpu())
te_loss = sum(loss_ls) / test_set.__len__()
te_pred = torch.cat(te_pred_ls, 0)
te_label = np.array(te_label_ls)
test_metrics = get_evaluation(te_label, te_pred.numpy(), list_metrics=["accuracy", "confusion_matrix"])
output_file.write(
"Epoch: {}/{} \nTest loss: {} Test accuracy: {} \nTest confusion matrix: \n{}\n\n".format(
epoch + 1, opt.num_epochs,
te_loss,
test_metrics["accuracy"],
test_metrics["confusion_matrix"]))
print("Epoch: {}/{}, Lr: {}, Loss: {}, Accuracy: {}".format(
epoch + 1,
opt.num_epochs,
optimizer.param_groups[0]['lr'],
te_loss, test_metrics["accuracy"]))
writer.add_scalar('Test/Loss', te_loss, epoch)
writer.add_scalar('Test/Accuracy', test_metrics["accuracy"], epoch)
model.train()
if te_loss + opt.es_min_delta < best_loss:
best_loss = te_loss
best_epoch = epoch
torch.save(model, opt.saved_path + os.sep + "whole_model_quickdraw")
if epoch - best_epoch > opt.es_patience > 0:
print("Stop training at epoch {}. The lowest loss achieved is {}".format(epoch, te_loss))
break
writer.close()
output_file.close()
if __name__ == "__main__":
opt = get_args()
train(opt)