General question

Question

General question

sainimohit23 opened this issue 5 years ago · comments

Hi Zheda,

I want to know that is this project based on this paper: https://arxiv.org/abs/1708.01715 and it's official implementation in pytorch?

I also want to know that have you implemented re-feeding?

Zheda (Marco) Mai · Answer 1 · Tue Jul 30 2019 03:37:18 GMT+0800 (China Standard Time)

Hi Zheda,

I want to know that is this project based on this paper: https://arxiv.org/abs/1708.01715 and it's official implementation in pytorch?

I also want to know that have you implemented re-feeding?

Hi, you are right and I also did some extra experiments based on the paper you mentioned. I did not implement re-feeding as I had a deadline to meet for this project.

Please let me know if you have the chance to implement re-feeding Keras. When I used the official code in pytorch in some other data, I found that re-feeding is very useful.

sainimohit23 · Answer 2 · Tue Jul 30 2019 15:30:45 GMT+0800 (China Standard Time)

@RaptorMai I couldn't find any way to implement re-feeding in keras. So, I have implemented it in tensorflow. I have a silly question about re-feeding that I have asked here: NVIDIA/DeepRecommender#29 . Can you answer that?

Zheda (Marco) Mai · Answer 3 · Wed Jul 31 2019 03:46:32 GMT+0800 (China Standard Time)

In dense re-feeding what do we show as output to model. For example let's say B is our training batch. For model input X and label y both will be equal to B. Let's say model gives y_hat as output for our batch B. Now, in dense re-feeding we give this y_hat as input to model. Now my question is what will be the labels in dense re-feeding? Will it be y_hat or is it B?

It should be B.