search

hexiangnan / adversarial_personalized_ranking

Adversarial Learning, Matrix Factorization, Recommendation
https://github.com/hexiangnan/adversarial_personalized_ranking
215 stars 62 forks source link

May I could know the configuration of your computer? Thanks! #9

Open HanXia001 opened 5 years ago

HanXia001 commented 5 years ago

Dear XiangNan, I have read your paper "Adversarial Personalized Ranking for Recommendation" more three times, and interested in the research which use GAN into IR.

I want to run the code at my computer, but my computer always report errors, such as "Reason:'MemoryError()' ".

So, May I could know the configuration of your computer? Thanks!

AaronHeee commented 5 years ago

@HanXia001 Hello, for your reference, our server environment is with Intel Xeon CPU E5-2630 @ 2.20 GHz and 64 GB memory.

AaronHeee commented 5 years ago

@HanXia001 Plus, we recommend your free memory is more than 8G, if you need to reproduce our experiments on Yelp.

HanXia001 commented 5 years ago

@HanXia001 Plus, we recommend your free memory is more than 8G, if you need to reproduce our experiments on Yelp.

First, thank you for the reply. The paper I have read many times, and every time there is harvest. As a graduate student, I also want to research on antagonistic learning. Yes, I need to reproduce experiments on Yelp, and there is problem: Memory Error. Maybe my computers' memory too low, all of the memory just 12G. I will consider improve it. At last, thank you very much kind.

AaronHeee commented 5 years ago

@HanXia001 Thanks for your reply. Maybe you can set the small value of cpu_count in multi-processing. Also, decreasing the batch size will be helpful to train these models with limited memory.

laugh12321 commented 5 years ago

Hello, when I run your code on Google Colab, there will be an error or disconnection after more than 10 hours. Is this related to the size of RAM storage (12 GB)?