File size: 14,467 Bytes
4057a1f
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
import torch
import torch.nn as nn
import numpy as np
import random
import os
from functools import partial
import warnings
from tqdm import tqdm
from torch.nn.init import trunc_normal_
import argparse
from optimizers import StableAdamW
from utils import evaluation_batch,WarmCosineScheduler, global_cosine_hm_adaptive, setup_seed, get_logger

# Dataset-Related Modules
from dataset import MVTecDataset, RealIADDataset
from dataset import get_data_transforms
from torchvision.datasets import ImageFolder
from torch.utils.data import DataLoader, ConcatDataset
from utils import augmentation

# Model-Related Modules
from models import vit_encoder
from models.uad import INP_Former
from models.vision_transformer import Mlp, Aggregation_Block, Prototype_Block


warnings.filterwarnings("ignore")
def main(args):
    # Fixing the Random Seed
    setup_seed(1)

    # Data Preparation
    data_transform, gt_transform = get_data_transforms(args.input_size, args.crop_size)

    if args.dataset == 'MVTec-AD' or args.dataset == 'VisA':
        train_data_list = []
        test_data_list = []
        for i, item in enumerate(args.item_list):
            train_path = os.path.join(args.data_path, item, 'train')
            test_path = os.path.join(args.data_path, item)

            train_data = ImageFolder(root=train_path, transform=data_transform)
            train_data.classes = item
            train_data.class_to_idx = {item: i}
            # Randomly select shot samples per class
            train_data.samples = [(sample[0], i) for sample in random.sample(train_data.samples, args.shot)]
            print_fn(train_data.samples)
            # Data augmentation
            augmented_data = []
            for img, label in train_data:
                augmented_imgs = augmentation(img)
                for aug_img in augmented_imgs:
                    augmented_data.append((aug_img, label))
            augmented_dataset = torch.utils.data.TensorDataset(
                torch.stack([item[0] for item in augmented_data]),
                torch.tensor([item[1] for item in augmented_data])
            )
            test_data = MVTecDataset(root=test_path, transform=data_transform, gt_transform=gt_transform, phase="test")
            train_data_list.append(augmented_dataset)
            test_data_list.append(test_data)
        train_data = ConcatDataset(train_data_list)
        train_dataloader = torch.utils.data.DataLoader(train_data, batch_size=args.batch_size, shuffle=True, num_workers=4, drop_last=True)
    elif args.dataset == 'Real-IAD' :
        train_data_list = []
        test_data_list = []
        for i, item in enumerate(args.item_list):
            train_data = RealIADDataset(root=args.data_path, category=item, transform=data_transform,
                                        gt_transform=gt_transform,
                                        phase='train')
            train_data.classes = item
            train_data.class_to_idx = {item: i}
            # Randomly select shot samples per class
            train_data.img_paths = [sample for sample in random.sample(list(train_data.img_paths), args.shot)]
            print_fn(train_data.img_paths)
            # Data augmentation
            augmented_data = []
            for img, label in train_data:
                augmented_imgs = augmentation(img)
                for aug_img in augmented_imgs:
                    augmented_data.append((aug_img, label))
            augmented_dataset = torch.utils.data.TensorDataset(
                torch.stack([item[0] for item in augmented_data]),
                torch.tensor([item[1] for item in augmented_data])
            )
            test_data = RealIADDataset(root=args.data_path, category=item, transform=data_transform,
                                       gt_transform=gt_transform,
                                       phase="test")
            train_data_list.append(augmented_dataset)
            test_data_list.append(test_data)

        train_data = ConcatDataset(train_data_list)
        train_dataloader = torch.utils.data.DataLoader(train_data, batch_size=args.batch_size, shuffle=True, num_workers=4,
                                                       drop_last=True)
    # Adopting a grouping-based reconstruction strategy similar to Dinomaly
    target_layers = [2, 3, 4, 5, 6, 7, 8, 9]
    fuse_layer_encoder = [[0, 1, 2, 3], [4, 5, 6, 7]]
    fuse_layer_decoder = [[0, 1, 2, 3], [4, 5, 6, 7]]

    # Encoder info
    encoder = vit_encoder.load(args.encoder)
    if 'small' in args.encoder:
        embed_dim, num_heads = 384, 6
    elif 'base' in args.encoder:
        embed_dim, num_heads = 768, 12
    elif 'large' in args.encoder:
        embed_dim, num_heads = 1024, 16
        target_layers = [4, 6, 8, 10, 12, 14, 16, 18]
    else:
        raise "Architecture not in small, base, large."

    # Model Preparation
    Bottleneck = []
    INP_Guided_Decoder = []
    INP_Extractor = []

    # bottleneck
    Bottleneck.append(Mlp(embed_dim, embed_dim * 4, embed_dim, drop=0.))
    Bottleneck = nn.ModuleList(Bottleneck)

    # INP
    INP = nn.ParameterList(
                    [nn.Parameter(torch.randn(args.INP_num, embed_dim))
                     for _ in range(1)])

    # INP Extractor
    for i in range(1):
        blk = Aggregation_Block(dim=embed_dim, num_heads=num_heads, mlp_ratio=4.,
                                qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-8))
        INP_Extractor.append(blk)
    INP_Extractor = nn.ModuleList(INP_Extractor)

    # INP_Guided_Decoder
    for i in range(8):
        blk = Prototype_Block(dim=embed_dim, num_heads=num_heads, mlp_ratio=4.,
                              qkv_bias=True, norm_layer=partial(nn.LayerNorm, eps=1e-8))
        INP_Guided_Decoder.append(blk)
    INP_Guided_Decoder = nn.ModuleList(INP_Guided_Decoder)

    model = INP_Former(encoder=encoder, bottleneck=Bottleneck, aggregation=INP_Extractor, decoder=INP_Guided_Decoder,
                             target_layers=target_layers,  remove_class_token=True, fuse_layer_encoder=fuse_layer_encoder,
                             fuse_layer_decoder=fuse_layer_decoder, prototype_token=INP)
    model = model.to(device)

    if args.phase == 'train':
        # Model Initialization
        trainable = nn.ModuleList([Bottleneck, INP_Guided_Decoder, INP_Extractor, INP])
        for m in trainable.modules():
            if isinstance(m, nn.Linear):
                trunc_normal_(m.weight, std=0.01, a=-0.03, b=0.03)
                if isinstance(m, nn.Linear) and m.bias is not None:
                    nn.init.constant_(m.bias, 0)
            elif isinstance(m, nn.LayerNorm):
                nn.init.constant_(m.bias, 0)
                nn.init.constant_(m.weight, 1.0)
        # define optimizer
        optimizer = StableAdamW([{'params': trainable.parameters()}],
                                lr=1e-3, betas=(0.9, 0.999), weight_decay=1e-4, amsgrad=True, eps=1e-10)
        lr_scheduler = WarmCosineScheduler(optimizer, base_value=1e-3, final_value=1e-4, total_iters=args.total_epochs*len(train_dataloader),
                                           warmup_iters=100)
        print_fn('train image number:{}'.format(len(train_data)))

        # Train
        for epoch in range(args.total_epochs):
            model.train()
            loss_list = []
            for img, _ in tqdm(train_dataloader, ncols=80):
                img = img.to(device)
                en, de, g_loss = model(img)
                loss = global_cosine_hm_adaptive(en, de, y=3)
                loss = loss + 0.2 * g_loss
                optimizer.zero_grad()
                loss.backward()
                nn.utils.clip_grad_norm(trainable.parameters(), max_norm=0.1)
                optimizer.step()
                loss_list.append(loss.item())
                lr_scheduler.step()
            print_fn('epoch [{}/{}], loss:{:.4f}'.format(epoch+1, args.total_epochs, np.mean(loss_list)))
            if (epoch + 1) % args.total_epochs == 0:
                auroc_sp_list, ap_sp_list, f1_sp_list = [], [], []
                auroc_px_list, ap_px_list, f1_px_list, aupro_px_list = [], [], [], []

                for item, test_data in zip(args.item_list, test_data_list):
                    test_dataloader = torch.utils.data.DataLoader(test_data, batch_size=args.batch_size, shuffle=False,
                                                                  num_workers=4)
                    results = evaluation_batch(model, test_dataloader, device, max_ratio=0.01, resize_mask=256)
                    auroc_sp, ap_sp, f1_sp, auroc_px, ap_px, f1_px, aupro_px = results
                    auroc_sp_list.append(auroc_sp)
                    ap_sp_list.append(ap_sp)
                    f1_sp_list.append(f1_sp)
                    auroc_px_list.append(auroc_px)
                    ap_px_list.append(ap_px)
                    f1_px_list.append(f1_px)
                    aupro_px_list.append(aupro_px)
                    print_fn(
                        '{}: I-Auroc:{:.4f}, I-AP:{:.4f}, I-F1:{:.4f}, P-AUROC:{:.4f}, P-AP:{:.4f}, P-F1:{:.4f}, P-AUPRO:{:.4f}'.format(
                            item, auroc_sp, ap_sp, f1_sp, auroc_px, ap_px, f1_px, aupro_px))

                print_fn('Mean: I-Auroc:{:.4f}, I-AP:{:.4f}, I-F1:{:.4f}, P-AUROC:{:.4f}, P-AP:{:.4f}, P-F1:{:.4f}, P-AUPRO:{:.4f}'.format(
                        np.mean(auroc_sp_list), np.mean(ap_sp_list), np.mean(f1_sp_list),
                        np.mean(auroc_px_list), np.mean(ap_px_list), np.mean(f1_px_list), np.mean(aupro_px_list)))
                torch.save(model.state_dict(), os.path.join(args.save_dir, args.save_name, 'model.pth'))
                model.train()
    elif args.phase == 'test':
        # Test
        model.load_state_dict(torch.load(os.path.join(args.save_dir, args.save_name, 'model.pth')), strict=True)
        auroc_sp_list, ap_sp_list, f1_sp_list = [], [], []
        auroc_px_list, ap_px_list, f1_px_list, aupro_px_list = [], [], [], []
        model.eval()
        for item, test_data in zip(args.item_list, test_data_list):
            test_dataloader = torch.utils.data.DataLoader(test_data, batch_size=args.batch_size, shuffle=False,
                                                          num_workers=4)
            results = evaluation_batch(model, test_dataloader, device, max_ratio=0.01, resize_mask=256)
            auroc_sp, ap_sp, f1_sp, auroc_px, ap_px, f1_px, aupro_px = results
            auroc_sp_list.append(auroc_sp)
            ap_sp_list.append(ap_sp)
            f1_sp_list.append(f1_sp)
            auroc_px_list.append(auroc_px)
            ap_px_list.append(ap_px)
            f1_px_list.append(f1_px)
            aupro_px_list.append(aupro_px)
            print_fn(
                '{}: I-Auroc:{:.4f}, I-AP:{:.4f}, I-F1:{:.4f}, P-AUROC:{:.4f}, P-AP:{:.4f}, P-F1:{:.4f}, P-AUPRO:{:.4f}'.format(
                    item, auroc_sp, ap_sp, f1_sp, auroc_px, ap_px, f1_px, aupro_px))

        print_fn(
            'Mean: I-Auroc:{:.4f}, I-AP:{:.4f}, I-F1:{:.4f}, P-AUROC:{:.4f}, P-AP:{:.4f}, P-F1:{:.4f}, P-AUPRO:{:.4f}'.format(
                np.mean(auroc_sp_list), np.mean(ap_sp_list), np.mean(f1_sp_list),
                np.mean(auroc_px_list), np.mean(ap_px_list), np.mean(f1_px_list), np.mean(aupro_px_list)))


if __name__ == '__main__':
    os.environ['CUDA_LAUNCH_BLOCKING'] = "1"
    parser = argparse.ArgumentParser(description='')

    # dataset info
    parser.add_argument('--dataset', type=str, default=r'MVTec-AD') # 'MVTec-AD' or 'VisA' or 'Real-IAD'
    parser.add_argument('--data_path', type=str, default=r'E:\IMSN-LW\dataset\mvtec_anomaly_detection')  # Replace it with your path.

    # save info
    parser.add_argument('--save_dir', type=str, default='./saved_results')
    parser.add_argument('--save_name', type=str, default='INP-Former-Few-Shot')

    # model info
    parser.add_argument('--encoder', type=str, default='dinov2reg_vit_base_14') # 'dinov2reg_vit_small_14' or 'dinov2reg_vit_base_14' or 'dinov2reg_vit_large_14'
    parser.add_argument('--input_size', type=int, default=448)
    parser.add_argument('--crop_size', type=int, default=392)
    parser.add_argument('--INP_num', type=int, default=6)

    # training info
    parser.add_argument('--total_epochs', type=int, default=200)
    parser.add_argument('--batch_size', type=int, default=16)
    parser.add_argument('--shot', type=int, default=4) # Number of samples
    parser.add_argument('--phase', type=str, default='train')

    args = parser.parse_args()
    args.save_name = args.save_name + f'-{args.shot}_dataset={args.dataset}_Encoder={args.encoder}_Resize={args.input_size}_Crop={args.crop_size}_INP_num={args.INP_num}'
    logger = get_logger(args.save_name, os.path.join(args.save_dir, args.save_name))
    print_fn = logger.info
    device = 'cuda:0' if torch.cuda.is_available() else 'cpu'

    # category info
    if args.dataset == 'MVTec-AD':
        # args.data_path = 'E:\IMSN-LW\dataset\mvtec_anomaly_detection' # '/path/to/dataset/MVTec-AD/'
        args.item_list = ['carpet', 'grid', 'leather', 'tile', 'wood', 'bottle', 'cable', 'capsule',
                 'hazelnut', 'metal_nut', 'pill', 'screw', 'toothbrush', 'transistor', 'zipper']
    elif args.dataset == 'VisA':
        # args.data_path = r'E:\IMSN-LW\dataset\VisA_pytorch\1cls'  # '/path/to/dataset/VisA/'
        args.item_list = ['candle', 'capsules', 'cashew', 'chewinggum', 'fryum', 'macaroni1', 'macaroni2',
                 'pcb1', 'pcb2', 'pcb3', 'pcb4', 'pipe_fryum']
    elif args.dataset == 'Real-IAD':
        # args.data_path = 'E:\IMSN-LW\dataset\Real-IAD'  # '/path/to/dataset/Real-IAD/'
        args.item_list = ['audiojack', 'bottle_cap', 'button_battery', 'end_cap', 'eraser', 'fire_hood',
                 'mint', 'mounts', 'pcb', 'phone_battery', 'plastic_nut', 'plastic_plug',
                 'porcelain_doll', 'regulator', 'rolled_strip_base', 'sim_card_set', 'switch', 'tape',
                 'terminalblock', 'toothbrush', 'toy', 'toy_brick', 'transistor1', 'usb',
                 'usb_adaptor', 'u_block', 'vcpill', 'wooden_beads', 'woodstick', 'zipper']

    main(args)