jin530 / loca Goto Github PK

Hi,

I am confused about how to aggregate global and local models. In the paper, there is a hyper-parameter alpha to weighted average outputs from global and local models. However, in the code, I did not find the place you use this weighted average method. Instead, it seems that you choose to use the output from the global model only when one user belongs to no local model. Like line 150 to 153 in LOCA_VAE.py. Am I understanding your code correctly?

BTW, thank you for the nice code and paper!

Hi, nice work!
But I found it is lacking validation dataset while test dataset as the measurement of early stop? Is it normal? Looking forword to your relpy~

Can you update the preprocessed datasets link so we can download the .rating datasets? Thank you so much.