Create dataset from a folder of images and a txt file with labels - image-processing

Question

I have two folders containing train and test images. And I have two files train.txt and test.txt containing the labels. I want to create a dataset for a PyTorch setting. I mean, I want to use DataLoader and I want the tuples (x_train, y_train) and (x_test, y_test). How can I do? Basically I think that in tensorflow I can use flow_from_dataframe, but in PyTorch?
As an example, in this image I show my pandas dataframe containing filenames and labels:
EDIT:
I am trying with a custom dataset class:
from torch.utils.data import Dataset, DataLoader
import os
import pandas as pd
from torchvision.io import read_image
class CoViDxDataset(Dataset):
def __init__(self, annotations_file, img_dir, transform=None, target_transform=None):
self.img_labels = annotations_file
self.img_dir = img_dir
self.transform = transform
self.target_transform = target_transform
def __len__(self):
return len(self.img_labels)
def __getitem__(self, idx):
img_path = os.path.join(self.img_dir, self.img_labels.iloc[idx, 0])
print(img_path)
image = read_image(img_path)
label = self.img_labels.iloc[idx, 1]
if self.transform:
image = self.transform(image)
if self.target_transform:
label = self.target_transform(label)
return image, label
train_dataset = CoViDxDataset(train_df, train_path)
When I do train_dataset[0] I have this error:
train/ARDSSevere.png
---------------------------------------------------------------------------
RuntimeError Traceback (most recent call last)
Input In [84], in <cell line: 1>()
----> 1 train_dataset[0]
Input In [82], in CoViDxDataset.__getitem__(self, idx)
17 img_path = os.path.join(self.img_dir, self.img_labels.iloc[idx, 0])
18 print(img_path)
---> 19 image = read_image(img_path)
20 label = self.img_labels.iloc[idx, 1]
21 if self.transform:
File ~/anaconda3/envs/openfl/lib/python3.8/site-packages/torchvision/io/image.py:245, in read_image(path, mode)
243 if not torch.jit.is_scripting() and not torch.jit.is_tracing():
244 _log_api_usage_once(read_image)
--> 245 data = read_file(path)
246 return decode_image(data, mode)
File ~/anaconda3/envs/openfl/lib/python3.8/site-packages/torchvision/io/image.py:47, in read_file(path)
45 if not torch.jit.is_scripting() and not torch.jit.is_tracing():
46 _log_api_usage_once(read_file)
---> 47 data = torch.ops.image.read_file(path)
48 return data
RuntimeError: Expected a non empty file
where train_path = 'train/' and train_df is the image below: