-
Notifications
You must be signed in to change notification settings - Fork 0
/
train.py
149 lines (103 loc) · 4.29 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
# %%
import torch
from torchvision import transforms
from torch.utils.data import DataLoader, random_split
from torchvision.datasets import ImageFolder
from torch.optim.lr_scheduler import StepLR
from sklearn.metrics import classification_report
import numpy as np
import pytorch_lightning as pl
from pytorch_lightning.callbacks.early_stopping import EarlyStopping
from model import MyModel
from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay
import matplotlib.pyplot as plt
import seaborn as sns
pl.seed_everything(2023)
GPU_IDX = 2
MAX_EPOCHS = 30
DETERMINISTIC =True
BATCH_SIZE = 64
LR = 0.001
# Learning rate scheduler params
STEP_SIZE = 20
GAMMA = 0.1
CLASSES = ['White', 'Black', 'Asian', 'Indian']
#%% Define transformations
train_transform = transforms.Compose([
transforms.Resize(size=128),
transforms.RandomCrop(104),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize( [0.5, 0.5, 0.5], [0.5, 0.5, 0.5])
])
test_transform = transforms.Compose([
transforms.Resize(size=128),
transforms.CenterCrop(size=104),
transforms.ToTensor(),
transforms.Normalize( [0.5, 0.5, 0.5],[0.5, 0.5, 0.5])
])
#%% Load datasets and apply transformations
train_dataset = ImageFolder('data/utk_races/train/', transform=train_transform)
val_dataset = ImageFolder('data/utk_races/val/', transform=test_transform)
test_dataset = ImageFolder('data/utk_races/test/', transform=test_transform)
#%% Create dataloaders
train_dataloader = DataLoader(train_dataset, batch_size=BATCH_SIZE, shuffle=True)
test_dataloader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=False)
val_dataloader = DataLoader(val_dataset, batch_size=BATCH_SIZE, shuffle=False)
device = torch.device(f"cuda:{GPU_IDX}") if torch.cuda.is_available() else torch.device("cpu")
#%% Model Training
model = MyModel(lr=LR, step_size=STEP_SIZE, gamma=GAMMA)
early_stopping_callback = EarlyStopping(monitor='val_loss', patience=20, mode='min')
trainer = pl.Trainer(devices=[GPU_IDX], accelerator='gpu', callbacks=[early_stopping_callback],
max_epochs=MAX_EPOCHS, deterministic=DETERMINISTIC)
trainer.fit(model, train_dataloader, val_dataloader)
#%% Evaluate model performance on test set
model.to(device)
model.eval()
correct = 0
total = 0
full_pred = torch.empty(0, 4).to(device)
full_label = torch.empty(0, dtype=torch.int32).to(device)
# Since we're not training, we don't need to calculate the gradients for our outputs
with torch.no_grad():
for data in test_dataloader:
images, labels = data[0].to(device), data[1].to(device)
# Compute outputs by running images through the network
outputs = model(images)
# The class with the highest probability is what we choose as prediction
_, predicted = torch.max(outputs.data, 1)
total += labels.size(0)
correct += (predicted == labels).sum().item()
# Store batches predictions & ground truth
full_pred = torch.cat((full_pred, outputs.data))
full_label = torch.cat((full_label, labels))
full_pred = full_pred.cpu()
full_label = full_label.cpu()
print("Accuracy of the network on the test set: {:.3f} %".format(100 * correct / total))
conf_matrix = confusion_matrix(full_label, torch.max(full_pred,1).indices)
fig, ax = plt.subplots(figsize=(8,6), dpi=100)
ax.set_title('Test confusion matrix')
# Show confusion matrix plot
display = ConfusionMatrixDisplay(conf_matrix, display_labels=CLASSES)
display.plot(ax=ax)
plt.savefig('visualizations/test_confusion_matrix.png')
# %% Show confusion matrix with percentage
# Normalize the confusion matrix by dividing each element by the sum of its row
conf_matrix_percentage = conf_matrix.astype('float') / conf_matrix.sum(axis=1)[:, np.newaxis]
plt.figure(figsize=(8, 6))
sns.heatmap(conf_matrix_percentage, annot=True, fmt=".3f", cmap="Blues",
xticklabels=CLASSES, yticklabels=CLASSES)
plt.xlabel("Predicted Label")
plt.ylabel("True Label")
plt.title("Confusion Matrix")
plt.savefig('visualizations/test_confusion_matrix_percentage.png')
plt.show()
# %% PLot val loss
flattened_loss = [x['val_loss'].cpu() for x in model.logged_metrics]
plt.plot(flattened_loss)
plt.xlabel('Index')
plt.ylabel('Loss')
plt.title('Validation Loss')
plt.savefig('visualizations/validation_loss.png')
plt.show()
# %%