Combine Train and Test Loader torchvision

I wish to use the complete data of MNSIT torchvision for training my Convolutional Autoencoder.

Is there any way to combine train_loader and test_loader ?

I’d probably look at instead of concatenating at the dataloader level (could be done using itertools.chain).
On a more abstract level, are you sure you don’t need a test set?

Best regards


Yes, actually I am doing pre-training for the model and I wish to use complete dataset during this part.

But I am not sure how to achieve this using ConcatDataset.

I am using MNIST dataset and I have loaded train_loader and test_loader using torchvision