Using WeightedRandomSampler in PyTorch

1

I need to implement a multi-label image classification model in PyTorch. However my data is not balanced, so I used the WeightedRandomSampler in PyTorch to create a custom dataloader. But when I iterate through the custom dataloader, I get the error : IndexError: list index out of range

Balanced Sampling between classes with torchvision DataLoader (Implemented the following code using this link)

def make_weights_for_balanced_classes(images, nclasses):                        
    count = [0] * nclasses                                                      
    for item in images:                                                         
        count[item[1]] += 1                                                     
    weight_per_class = [0.] * nclasses                                      
    N = float(sum(count))                                                   
    for i in range(nclasses):                                                   
        weight_per_class[i] = N/float(count[i])                                 
    weight = [0] * len(images)                                              
    for idx, val in enumerate(images):                                          
        weight[idx] = weight_per_class[val[1]]                                  
    return weight 

weights = make_weights_for_balanced_classes(train_dataset.imgs, len(full_dataset.classes))
weights = torch.DoubleTensor(weights)
sampler = WeightedRandomSampler(weights, len(weights))

train_loader = DataLoader(train_dataset, batch_size=4,sampler = sampler, pin_memory=True)

Thanks a lot in advance !

2

Which line of code is throwing this error?
The code looks generally alright and I cannot spot an obvious mistake.