Wrapper for a PyTorch classifier which allows it to output prediction sets. The sets are theoretically guaranteed to contain the true class with high probability (via conformal prediction).
Why do you think that https://github.com/awslabs/fortuna is a good alternative to conformal_classification