sigmoid is missing in QuestionAttnGRU.py

Question

sigmoid is missing in QuestionAttnGRU.py

keithhans opened this issue 7 years ago · 5 comments

keithhans commented 7 years ago

According to equation 6 in the paper, there should be a sigmoid on K.dot(GRU_inputs, W_g1).

Answer 1 · 2017-10-07T13:45:57.000Z

Completely agree with you. Will be fixed soon.

Answer 2 · 2017-10-07T16:01:07.000Z

Just noticed it is also missing in SelfAttnGRU.py

Answer 3 · 2017-10-07T17:02:54.000Z

We also noticed that.
Sigmoid disappeared after we switched from using Dense layers (with sigmoid activation) to SharedWeight layers.
Now I'm trying to repeat training process. We hope can get better scores after fixing this issue.

Answer 4 · 2017-10-25T17:10:17.000Z

Pull request is merged.
Not closing the issue until we compare new scores.
@MartinXPN please rerun the instruction steps from README.md

Answer 5 · 2017-11-05T06:00:11.000Z

After repeating all the steps in readme the performance didn't change much. Accuracy reached up to 60% (a bit less)