Problem

This is a valuable source of learning materials when trying to deeply understand the mechanism of several neural network architectures. But I notice that you have not updated the repo since 2017.

Do you plan to make any update in the coming future?

In Keras I have seen load_model and model.predict_classes for loading a previously trained model and testing it. How can I achieve this in numpydl?

Thank you for a great framework.
In the current implementation, I do not find an option to set padding setting for a convolutional layer (options would be 'valid', 'same', ...). Am I right that there is no an option to set it?
If not I am very interested in having it. So I would implement it if you are interested in this option. Just let me know if you are interested and I will prepare a pull request.

NumpyDL haven't sequential model.