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
| from torchvision import datasets, transforms
from torch.utils.data import DataLoader
import torch.nn as nn
import torch.optim as optim
import torch
from pathlib import Path
import matplotlib.pyplot as plt
def get_dataloader():
transform = transforms.ToTensor()
train_dataset = datasets.FashionMNIST(
root="./datasets",
train=True,
download=True,
transform=transform,
)
train_loader = DataLoader(
train_dataset,
batch_size=128,
shuffle=True,
)
test_dataset = datasets.FashionMNIST(
root="./datasets",
train=False,
download=True,
transform=transform,
)
test_loader = DataLoader(
test_dataset,
batch_size=128,
shuffle=False,
)
return train_loader, test_loader
class MLP(nn.Module):
def __init__(self):
super().__init__()
self.flatten = nn.Flatten()
self.linear1 = nn.Linear(28*28, 256)
self.relu = nn.ReLU()
self.linear2 = nn.Linear(256, 10)
def forward(self, x):
x = self.flatten(x)
x = self.linear1(x)
x = self.relu(x)
x = self.linear2(x)
return x
def main():
train_loader, test_loader = get_dataloader()
model = MLP()
criterion = nn.CrossEntropyLoss()
optimizer = optim.Adam(model.parameters(), lr=0.001)
epochs = 5
path = Path("checkpoints")
path.mkdir(exist_ok=True)
path = Path("plots")
path.mkdir(exist_ok=True)
train_losses = []
val_losses = []
best_val_acc = 0.0
for epoch in range(epochs):
model.train()
train_loss = 0.0
train_correct = 0
train_total = 0
for images, labels in train_loader:
output = model(images)
preds = output.argmax(dim=1)
train_correct += (preds == labels).sum().item()
train_total += labels.size(0)
loss = criterion(output, labels)
train_loss += loss.item() * labels.size(0)
optimizer.zero_grad()
loss.backward()
optimizer.step()
train_loss = train_loss / train_total
train_acc = train_correct / train_total
model.eval()
val_loss = 0.0
val_correct = 0
val_total = 0
with torch.no_grad():
for images, labels in test_loader:
output = model(images)
preds = output.argmax(dim=1)
val_correct += (preds == labels).sum().item()
val_total += labels.size(0)
loss = criterion(output, labels)
val_loss += loss.item() * labels.size(0)
val_loss = val_loss / val_total
val_acc = val_correct / val_total
train_losses.append(train_loss)
val_losses.append(val_loss)
print(f"Epoch {epoch+1}/{epochs}")
print(f"Train Loss: {train_loss:.4f}\nTrain Acc: {train_acc:.4f}")
print(f"Val Loss: {val_loss:.4f}\nVal Acc: {val_acc:.4f}\n")
if val_acc > best_val_acc:
best_val_acc = val_acc
torch.save({
"model_state_dict": model.state_dict(),
"optimizer_state_dict": optimizer.state_dict(),
"epoch": epoch+1,
"val_acc": val_acc,
}, "./checkpoints/best_model.pth")
epochs_range = range(1, epochs + 1)
plt.figure(figsize=(10, 5))
plt.plot(epochs_range, train_losses, label="Train Loss")
plt.plot(epochs_range, val_losses, label="Val Loss")
plt.xlabel("Epoch")
plt.ylabel("Loss")
plt.title("Train and Val Loss")
plt.legend()
plt.savefig("./plots/loss_plot.png")
plt.close()
if __name__ == "__main__":
main()
|
