pydeeplearn
pydeeplearn copied to clipboard
serial validation set
tell me your opinion please, it can be done for other neural netoworks (or there could be some abstraction for common methods :) )
There is a problem with this code. The data is no longer shuffled if there is no validation performed. It is important not to change that.
Please add another method in which the training is deterministic (I guess that is what you are looking for).
ah, ofc! should be fixed now :)