List index out of range

[KanZa]

I am using a custom dataset that has the following dimensions (3, 64,64) there are 28 classes. The code works absolutely fine with pre-trained ResNet50 for cifar10 and a smaller dataset with 4 classes with similar dimensions. but for 28 classes after some cycles, it gives the following error for ‘ct’ in data loader in def getitem(self, index):

import torch 
import torchvision 
from torch.utils.data import Dataset
import torchvision.transforms as T
from torchvision.datasets import CIFAR10, ImageFolder
from config import *

class MyDataset(Dataset):
    def __init__(self, dataset_name, train_flag, transf):
        self.dataset_name = dataset_name
        if self.dataset_name == "cifar10":
            self.cifar10 = CIFAR10('../cifar10', train=train_flag, 
                                    download=True, transform=transf)
        if self.dataset_name == "ct":
            self.ct = ImageFolder(root='/Dataset/radiology_ai/CT/Split-CT-abd/train', transform=transf)
        if self.dataset_name == "satellite":
            self.satellite = ImageFolder(root='/Dataset/Satellite/train', transform=transf)
        
    def __getitem__(self, index):
        if self.dataset_name == "cifar10":
            data, target = self.cifar10[index]
      
        if self.dataset_name == "ct":
            data, target = self.ct[index]
              
        if self.dataset_name == "satellite":

            data, target = self.satellite[index]
            
        return data, target, index

    def __len__(self):
        if self.dataset_name == "cifar10":
            return len(self.cifar10)
        
        if self.dataset_name == "ct":
            return len(self.ct)
        
        if self.dataset_name == "satellite":
            return len(self.ct)

def load_dataset(dataset):
    train_transform = T.Compose([
        T.RandomHorizontalFlip(),
        T.RandomCrop(size=64, padding=4),
        T.ToTensor(),
        T.Normalize([0.4914, 0.4822, 0.4465], [0.2023, 0.1994, 0.2010]) # T.Normalize((0.5071, 0.4867, 0.4408), (0.2675, 0.2565, 0.2761)) # CIFAR-100
    ])

    test_transform = T.Compose([
        T.ToTensor(),
        T.Normalize([0.4914, 0.4822, 0.4465], [0.2023, 0.1994, 0.2010]) # T.Normalize((0.5071, 0.4867, 0.4408), (0.2675, 0.2565, 0.2761)) # CIFAR-100
    ])


    if dataset == 'cifar10': 
        data_train = CIFAR10('../cifar10', train=True, download=True, transform=train_transform)
        data_unlabeled = MyDataset(dataset, True, test_transform)
        data_test  = CIFAR10('../cifar10', train=False, download=True, transform=test_transform)
        NO_CLASSES = 10
        adden = ADDENDUM
        no_train = NUM_TRAIN
    
    if dataset == 'ct': 
        data_train = ImageFolder(root='/Dataset/radiology_ai/CT/Split-CT-abd/train', transform=train_transform)
        data_unlabeled = MyDataset(dataset, True, test_transform)
        data_test  = ImageFolder(root='/Dataset/radiology_ai/CT/Split-CT-abd/val', transform=test_transform)
        NO_CLASSES = 28
        adden = ADDENDUM
        no_train = NUM_TRAIN
    
    if dataset == 'satellite': 
        data_train = ImageFolder(root='/Dataset/Satellite/train', transform=train_transform)
        data_unlabeled = MyDataset(dataset, True, test_transform)
        data_test  = ImageFolder(root='/Dataset/Satellite/test', transform=test_transform)
        NO_CLASSES = 4
        adden = ADDENDUM
        no_train = NUM_TRAIN
    return data_train, data_unlabeled, data_test, adden, NO_CLASSES, no_train

[ptrblck]

Check the stacktrace to see which operation fails and then check the length of the list as well as the index. This should give you more information about the out-of-bounds indexing.

[KanZa]

The code debugger stops here

if self.dataset_name == "ct":
            data, target = self.ct[index]
            print("Data:", data.size())
            print("Target:", target)