ExampleLoadingOfDataset
Browse files- DatasetGenerator.py +48 -0
DatasetGenerator.py
ADDED
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
### Example loading of the dataset
|
2 |
+
|
3 |
+
import torch
|
4 |
+
from torch.utils.data import Dataset
|
5 |
+
from torchvision import datasets
|
6 |
+
from torchvision.transforms import ToTensor
|
7 |
+
import matplotlib.pyplot as plt
|
8 |
+
import zipfile
|
9 |
+
import os
|
10 |
+
import pandas as pd
|
11 |
+
from torchvision.io import read_image
|
12 |
+
|
13 |
+
class CustomImageDataset(Dataset):
|
14 |
+
def __init__(self, annotations_file, img_dir, transform=None, target_transform=None):
|
15 |
+
self.img_labels = pd.read_csv(annotations_file)
|
16 |
+
self.img_dir = img_dir
|
17 |
+
self.transform = transform
|
18 |
+
self.target_transform = target_transform
|
19 |
+
|
20 |
+
def __len__(self):
|
21 |
+
return len(self.img_labels)
|
22 |
+
|
23 |
+
def __getitem__(self, idx):
|
24 |
+
img_path = os.path.join(self.img_dir, self.img_labels.iloc[idx, -1])
|
25 |
+
image = read_image(img_path)
|
26 |
+
label = self.img_labels.iloc[idx, 2]
|
27 |
+
if self.transform:
|
28 |
+
image = self.transform(image)
|
29 |
+
if self.target_transform:
|
30 |
+
label = self.target_transform(label)
|
31 |
+
return image, label
|
32 |
+
|
33 |
+
with zipfile.ZipFile("150_Dataset(1).zip", 'r') as zip_ref:
|
34 |
+
zip_ref.extractall(".")
|
35 |
+
|
36 |
+
train_dataset = CustomImageDataset(annotations_file="./images/train/train.csv",
|
37 |
+
img_dir="./images/train")
|
38 |
+
|
39 |
+
train_dataloader = DataLoader(train_dataset, batch_size=12, shuffle=True)
|
40 |
+
|
41 |
+
train_features, train_labels = next(iter(train_dataloader))
|
42 |
+
print(f"Feature batch shape: {train_features.size()}")
|
43 |
+
print(f"Labels batch shape: {len(train_labels)}")
|
44 |
+
img = train_features[0].squeeze()
|
45 |
+
label = train_labels[0]
|
46 |
+
plt.imshow(img, cmap="gray")
|
47 |
+
plt.show()
|
48 |
+
print(f"Label: {label}")
|