Lower r1, r5, r10 and MAP results

[heerduo]

Thank for your work. I set batchz-size is 100 or 184, and lr is 2e-5 or 1.414e-4, I got r1 0.000, r5 0.148 r10 0.208, map 5.395. And I have load the pre-training model.

[mokizzz]

Mine are better than you. I got r1 30, map 15, and train from scratch. (batch_size=96, lr=2e-4, Duke->Market)

[KovenYu]

Sorry for the late reply. What if using the default setting? If that's okay, could you please modify one parameter at a time, so that we can locate where the problem is? Also, did you observe NAN loss like this issue?

[DeepLearner415]

Hey I tried to run the demo code, it kept saying that 'cant find data/ml_Market.dat', I made sure I downloaded all the datasets and other files and saved in the correct directory. any advice ? Thanks