How to deal with unequal length input in CNN or MLP?

vision
8

@beann You can also make your custom collate function for making a batch of variable sizes.
see this How to create a dataloader with variable-size input
And then you can use adaptive pooling inside the model.