Difference between FM and AFM is not as pronounced as in the original paper about attentional_factorization_machine HOT 3 OPEN

As mentioned in one of the github issues, we also tried to set batch_size to 2048 and lamda_attention to 32 but with no success. FM and AFM are almost identical: FM RMSE is 0.47 and AFM RMSE is 0.471.

from attentional_factorization_machine.

As mentioned in one of the github issues, we also tried to set batch_size to 2048 and lamda_attention to 32 but with no success. FM and AFM are almost identical: FM RMSE is 0.47 and AFM RMSE is 0.471.

I had the same problem, how do you solve it finally?

from attentional_factorization_machine.

As mentioned in one of the github issues, we also tried to set batch_size to 2048 and lamda_attention to 32 but with no success. FM and AFM are almost identical: FM RMSE is 0.47 and AFM RMSE is 0.471.

I had the same problem, how do you solve it finally?

We haven't solved it yet. If you find the solution we would be very interested in it.

from attentional_factorization_machine.