How to input more than one image into a model

I have images in a folder called 2. I want to train a model to output images from folder 3 when input images from folder 2.

For this, to work I have to train in batches or train on the whole set at once!

But how can I input more than one image into the model to train???

I looked everywhere and IDK how-to! I’ve tried making a tensor of tensors of the full data and doing
model(data) but uh oh it doesn’t work cuz it’s not the right dimensions.

The MNIST example gives a relatively short example of batched training:
examples/ at master · pytorch/examples (, primarily through the use of a DataLoader: — PyTorch 1.9.1 documentation.

what is data in the train function??