search

brightmart / text_classification

all kinds of text classification models and more with deep learning
MIT License
7.83k stars 2.57k forks source link

keep the drop_out_prob the same as in training #57

Closed acadTags closed 6 years ago

acadTags commented 6 years ago

In training the drop_out_prob is 0.5. This setting should be the same in evaluation.

acadTags commented 6 years ago

I realise that I was wrong: the dropout should only be kept in training, not in testing.