-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathfine_tuning_resnet.py
More file actions
250 lines (191 loc) · 8.25 KB
/
fine_tuning_resnet.py
File metadata and controls
250 lines (191 loc) · 8.25 KB
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
#### Ce fichier génère des embedding pour les images de train et de test en utilisant un modèle pré-entrainé ResNet50 fine tuné à prédire les classes de vêtement
import os
import pandas as pd
import numpy as np
import torch
from torchvision import models, transforms
from torch.utils.data import DataLoader, Dataset, random_split
from PIL import Image
import torch.nn as nn
import torch.optim as optim
from torch.optim.lr_scheduler import ReduceLROnPlateau
# Device
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print(f"Using: {device}")
# Paths
csv_file = "product_list.csv"
images_folder = "DAM_modified"
# Load labels
product_data = pd.read_csv(csv_file)
id_to_label = dict(zip(product_data['MMC'], product_data['Product_BusinessUnitDesc']))
unique_labels = list(set(product_data['Product_BusinessUnitDesc']))
label_to_idx = {label: idx for idx, label in enumerate(unique_labels)}
# Data augmentation for training
train_transform = transforms.Compose([
transforms.Resize((256, 256)),
transforms.RandomHorizontalFlip(),
transforms.RandomRotation(10),
transforms.ColorJitter(brightness=0.2, contrast=0.2),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
])
# Simple transforms for validation
val_transform = transforms.Compose([
transforms.Resize((256, 256)),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
])
class CustomDataset(Dataset):
def __init__(self, images_folder, id_to_label, transform, label_to_idx):
self.images_folder = images_folder
self.id_to_label = id_to_label
self.transform = transform
self.label_to_idx = label_to_idx
self.image_files = [f for f in os.listdir(images_folder) if f.split('.')[0] in id_to_label]
print(f"Total images: {len(self.image_files)}")
def __getitem__(self, index):
image_file = self.image_files[index]
image_path = os.path.join(self.images_folder, image_file)
image_id = image_file.split('.')[0]
label = self.id_to_label[image_id]
image = Image.open(image_path).convert('RGB')
if self.transform:
image = self.transform(image)
label_idx = self.label_to_idx[label]
return image, label_idx
def __len__(self):
return len(self.image_files)
# Create datasets with different transforms
full_dataset = CustomDataset(images_folder, id_to_label, train_transform, label_to_idx)
# Split into train and validation
train_size = int(0.8 * len(full_dataset))
val_size = len(full_dataset) - train_size
train_dataset, val_dataset = random_split(full_dataset, [train_size, val_size])
# Override transform for validation dataset
val_dataset.dataset.transform = val_transform
# Create dataloaders
train_loader = DataLoader(train_dataset, batch_size=16, shuffle=True, num_workers=4, pin_memory=True)
val_loader = DataLoader(val_dataset, batch_size=16, shuffle=False, num_workers=4, pin_memory=True)
# Model
model = models.resnet50(weights=models.ResNet50_Weights.IMAGENET1K_V1)
num_classes = len(unique_labels)
model.fc = nn.Linear(model.fc.in_features, num_classes)
# Progressive unfreezing
def unfreeze_layer(model, layer_name):
for name, param in model.named_parameters():
if layer_name in name:
param.requires_grad = True
# Initially freeze all layers except fc
for param in model.parameters():
param.requires_grad = False
for param in model.fc.parameters():
param.requires_grad = True
model = model.to(device)
# Training settings
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(filter(lambda p: p.requires_grad, model.parameters()), lr=0.00001)
scheduler = ReduceLROnPlateau(optimizer, mode='min', factor=0.1, patience=2, verbose=True)
def validate(model, val_loader, criterion):
model.eval()
val_loss = 0
correct = 0
total = 0
with torch.no_grad():
for images, labels in val_loader:
images, labels = images.to(device), labels.to(device)
outputs = model(images)
loss = criterion(outputs, labels)
val_loss += loss.item()
_, predicted = torch.max(outputs, 1)
total += labels.size(0)
correct += (predicted == labels).sum().item()
return val_loss / len(val_loader), 100 * correct / total
# Training loop with progressive unfreezing
print("Starting training...")
best_val_acc = 0
patience = 10
no_improve = 0
def training(epochs_per_stage):
global best_val_acc # Déclare best_val_acc comme une variable globale
global patience
global no_improve
for epoch in range(epochs_per_stage):
model.train()
train_loss = 0
correct = 0
total = 0
for images, labels in train_loader:
images, labels = images.to(device), labels.to(device)
optimizer.zero_grad()
outputs = model(images)
loss = criterion(outputs, labels)
loss.backward()
optimizer.step()
train_loss += loss.item()
_, predicted = torch.max(outputs, 1)
total += labels.size(0)
correct += (predicted == labels).sum().item()
train_loss = train_loss / len(train_loader)
train_acc = 100 * correct / total
# Validation
val_loss, val_acc = validate(model, val_loader, criterion)
print(f"Epoch [{epoch + 1}/{epochs_per_stage}]")
print(f"Train Loss: {train_loss:.4f}, Train Acc: {train_acc:.2f}%")
print(f"Val Loss: {val_loss:.4f}, Val Acc: {val_acc:.2f}%")
scheduler.step(val_loss)
if val_acc > best_val_acc:
best_val_acc = val_acc
no_improve = 0
torch.save(model.state_dict(), 'best_model.pth')
else:
no_improve += 1
if no_improve >= patience:
print("Early stopping triggered")
break
# Stage 1: Only fc layer
print("Stage 1: Training fc layer")
training(10)
# Stage 2: Unfreeze layer4
print("Stage 2: Training layer4")
unfreeze_layer(model, 'layer4')
optimizer = optim.Adam(filter(lambda p: p.requires_grad, model.parameters()), lr=0.00005)
scheduler = ReduceLROnPlateau(optimizer, mode='min', factor=0.1, patience=3, verbose=True)
# Repeat training loop for layer4 (same as above)
training(6)
# # Stage 3: Unfreeze layer3
# print("Stage 3: Training layer3")
# unfreeze_layer(model, 'layer3')
# optimizer = optim.Adam(filter(lambda p: p.requires_grad, model.parameters()), lr=0.00005)
# scheduler = ReduceLROnPlateau(optimizer, mode='min', factor=0.1, patience=3, verbose=True)
# # Repeat training loop for layer3
# training(5)
print("Training completed.")
# Load best model for embedding extraction
model.load_state_dict(torch.load('best_model.pth'))
embedding_model = nn.Sequential(*list(model.children())[:-1])
embedding_model.eval()
def extract_embeddings(image_folder, transform, embedding_model):
embeddings = {}
image_files = [file for file in os.listdir(image_folder) if file.endswith((".png", ".jpg", ".jpeg"))]
total_files = len(image_files)
print(f"Total images to process: {total_files}")
with torch.no_grad():
for i, file in enumerate(image_files):
image_path = os.path.join(image_folder, file)
image = Image.open(image_path).convert("RGB")
input_tensor = transform(image).unsqueeze(0).to(device)
embedding = embedding_model(input_tensor)
embedding = embedding.squeeze().cpu().numpy()
embeddings[os.path.splitext(file)[0]] = embedding
if (i + 1) % 100 == 0:
print(f"Progress: {(i+1)/total_files*100:.2f}% ({i+1}/{total_files})")
return embeddings
# Extract embeddings using the best model
print("Extracting embeddings...")
train_folder = './DAM_modified'
test_folder = './test_image_modified'
embeddings_train = extract_embeddings(train_folder, val_transform, embedding_model)
embeddings_test = extract_embeddings(test_folder, val_transform, embedding_model)
np.save("embeddings_train_modified_3.npy", embeddings_train)
np.save("embeddings_test_modified_3.npy", embeddings_test)
print("Embeddings extracted and saved.")