TransUnet/trainer1.py at master · MacDunno/TransUnet · GitHub

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
158
159
160
161
162
163
import argparse
import logging
import os
import random
import sys
import time
import numpy as np
import torch
import torch.nn as nn
import torch.optim as optim
from tensorboardX import SummaryWriter
from torch.nn.modules.loss import CrossEntropyLoss
from torch.utils.data import DataLoader
from tqdm import tqdm
from utils import DiceLoss
from torchvision import transforms
from medpy import metric

def trainer_synapse(args, model, snapshot_path):
    from datasets.dataset_synapse import Synapse_dataset, RandomGenerator, RandomGenerator1
    logging.basicConfig(filename=snapshot_path + "/log.txt", level=logging.INFO,
                        format='[%(asctime)s.%(msecs)03d] %(message)s', datefmt='%H:%M:%S')
    logging.getLogger().addHandler(logging.StreamHandler(sys.stdout))
    logging.info(str(args))
    base_lr = args.base_lr
    num_classes = args.num_classes
    batch_size = args.batch_size * args.n_gpu
    img_path = args.root_path+"/img"
    label_path = args.root_path+"/label"
    db_train = Synapse_dataset(img_dir=img_path, label_dir=label_path, list_dir=args.list_dir, split="train",
                               transform=transforms.Compose(
                                    [RandomGenerator(output_size=[args.img_size, args.img_size])]))
    val_image_path = "data/Synapse/less002/val/img"
    val_label_path = "data/Synapse/less002/val/label"
    val_dir = "lists/lists_Synapse/less002"
    db_val = Synapse_dataset(img_dir=val_image_path, label_dir=val_label_path, list_dir=val_dir, split="val",
                             transform=transforms.Compose(
                                    [RandomGenerator1(output_size=[args.img_size, args.img_size])]))
    print("The length of train set is: {}".format(len(db_train)))

    def worker_init_fn(worker_id):
        random.seed(args.seed + worker_id)

    trainloader = DataLoader(db_train, batch_size=batch_size, shuffle=True, num_workers=8, pin_memory=True,
                             worker_init_fn=worker_init_fn)
    valloader = DataLoader(db_val, batch_size=batch_size, shuffle=False, num_workers=8, pin_memory=True)
    if args.n_gpu > 1:
        model = nn.DataParallel(model)
    model.train()
    ce_loss = CrossEntropyLoss()
    dice_loss = DiceLoss(num_classes)
    # optimizer = optim.Adam(model.parameters(), lr=base_lr, betas=(0.9, 0.999), eps=1e-08, weight_decay=0.0001, amsgrad=False)
    optimizer = optim.SGD(model.parameters(), lr=base_lr, momentum=0.9, weight_decay=0.0001)
    writer = SummaryWriter(snapshot_path + '/log')
    iter_num = 0
    max_epoch = args.max_epochs
    max_iterations = args.max_epochs * len(trainloader)  # max_epoch = max_iterations // len(trainloader) + 1
    logging.info("{} iterations per epoch. {} max iterations ".format(len(trainloader), max_iterations))
    best_performance = 0.0

    best_loss = float('inf')
    best_dice = 0.0
    best_model_path = os.path.join(snapshot_path, 'best_model.pth')

    def validate(model, valloader, criterion_ce, criterion_dice):
        model.eval()
        val_loss = 0
        val_loss_ce = 0
        val_loss_dice = 0
        total_dice = 0

        with torch.no_grad():
            for batch in valloader:
                images, labels = batch['image'].cuda(), batch['label'].cuda()
                outputs = model(images)
                loss_ce = criterion_ce(outputs, labels.long())
                loss_dice = criterion_dice(outputs, labels, softmax=True)
                loss = 0.5 * loss_ce + 0.5 * loss_dice

                val_loss += loss.item()
                val_loss_ce += loss_ce.item()
                val_loss_dice += loss_dice.item()
                outputs = torch.softmax(outputs, dim=1)
                pred = outputs.argmax(dim=1).cpu().numpy()
                true = labels.cpu().numpy()

                dice_score = metric.binary.dc(pred==1, true==1)
                total_dice += dice_score

        avg_val_loss = val_loss / len(valloader)
        avg_val_loss_ce = val_loss_ce / len(valloader)
        avg_val_loss_dice = val_loss_dice / len(valloader)
        avg_val_dice = total_dice / len(valloader)

        model.train()
        return avg_val_loss, avg_val_loss_ce, avg_val_loss_dice, avg_val_dice

    for epoch_num in range(max_epoch):
        epoch_loss = 0
        epoch_loss_ce = 0
        epoch_loss_dice = 0
        total_dice = 0

        for i_batch, sampled_batch in enumerate(trainloader):
            image_batch, label_batch = sampled_batch['image'], sampled_batch['label']
            image_batch, label_batch = image_batch.cuda(), label_batch.cuda()
            outputs = model(image_batch)
            loss_ce = ce_loss(outputs, label_batch[:].long())
            loss_dice = dice_loss(outputs, label_batch, softmax=True)
            loss = 0.5 * loss_ce + 0.5 * loss_dice

            optimizer.zero_grad()
            loss.backward()
            optimizer.step()

            epoch_loss += loss.item()
            epoch_loss_ce += loss_ce.item()
            epoch_loss_dice += loss_dice.item()

            with torch.no_grad():
                outputs = torch.softmax(outputs, dim=1)
                pred = outputs.argmax(dim=1).cpu().numpy()
                true = label_batch.cpu().numpy()
                dice_score = metric.binary.dc(pred == 1, true == 1)  # 假设1是目标类别
                total_dice += dice_score

            lr_ = base_lr * (1.0 - iter_num / max_iterations) ** 0.9
            for param_group in optimizer.param_groups:
                param_group['lr'] = lr_

            iter_num = iter_num + 1
            writer.add_scalar('info/lr', lr_, iter_num)
            writer.add_scalar('info/total_loss', loss, iter_num)
            writer.add_scalar('info/loss_ce', loss_ce, iter_num)

        avg_loss = epoch_loss / len(trainloader)
        avg_loss_ce = epoch_loss_ce / len(trainloader)
        avg_loss_dice = epoch_loss_dice / len(trainloader)
        avg_dice = total_dice / len(trainloader)

        # 验证
        avg_val_loss, avg_val_loss_ce, avg_val_loss_dice, avg_val_dice = validate(model, valloader, ce_loss, dice_loss)

        logging.info(f'Epoch [{epoch_num+1}/{max_epoch}] - '
                     f'Train Loss: {avg_loss:.4f}, Val Loss: {avg_val_loss:.4f}, '
                     f'Train Dice: {avg_dice:.4f}, Val Dice: {avg_val_dice:.4f}')

        writer.add_scalar('epoch/avg_loss', avg_loss, epoch_num)
        writer.add_scalar('epoch/avg_loss_ce', avg_loss_ce, epoch_num)
        writer.add_scalar('epoch/avg_loss_dice', avg_loss_dice, epoch_num)
        writer.add_scalar('epoch/avg_dice', avg_dice, epoch_num)
        writer.add_scalar('epoch/val_loss', avg_val_loss, epoch_num)
        writer.add_scalar('epoch/val_dice', avg_val_dice, epoch_num)

        if avg_val_dice > best_dice:
            best_dice = avg_val_dice
            best_loss = avg_val_loss
            torch.save(model.state_dict(), best_model_path)
            logging.info(f"New best model saved with loss: {best_loss:.4f} and Dice: {best_dice:.4f}")

    writer.close()
    logging.info(f"Best model saved with loss: {best_loss:.4f} and Dice: {best_dice:.4f}")
    return "Training Finished!"