Try to use the Oversampling callback in a Multi-label image classification problem. When use data.train_dl.dl.batch_sampler = BatchSampler(WeightedRandomSampler(weights,total_len_oversample), data.train_dl.batch_size,False)
Raise the error: ValueError: batch_sampler attribute should not be set after DataLoader is initialized.
it seems everything is the same as in "Oversampling MNIST with fastai "kernel. I tried to find the reason and not working…Thank you so much for your help!!!