When the #hypefollowers ask me for my opinion on their #deeplearning manuscript... For an application that allegedly outperforms everything and the text is full of

> both training and evaluation were performed over the training set because of the lack of data for the test set

🤦‍♂️