A unified framework of perturbation and gradient-based attribution methods for Deep Neural Networks interpretability. DeepExplain also includes support for Shapley Values sampling. (ICLR 2018)
Hi Marco,
Thank you very much for developing and publishing this very helpful and well-written library!
I'm thinking about replacing print-statements with proper logging.
An advantage would be that the logged information could be easily discarded on demand.
Do you see any disadvantages?
Hi Marco, Thank you very much for developing and publishing this very helpful and well-written library! I'm thinking about replacing print-statements with proper logging. An advantage would be that the logged information could be easily discarded on demand. Do you see any disadvantages?