Implement about code


I’m a student who studies deep learning with Pytorch.

I have a question, but I want to learn only a specific character dataset (for example, alphabet V) when learning EMNIST dataset with Pytorch, but I want to know how to implement this with code.


You could manipulate the targets such that your target class uses the class index 1 while all other classes are assigned the class index 0.
Afterwards you can treat it as a binary classification and could use e.g. nn.BCEWithLogitsLoss as your criterion.