SASRec.pytorch icon indicating copy to clipboard operation
SASRec.pytorch copied to clipboard
verified publisher icon pmixer

training not converge on Amazon Beauty dataset according to the paper's hyperparameter setting

Open lipingcoding opened this issue 5 years ago • 3 comments

lipingcoding avatar Nov 13 '20 00:11 lipingcoding

@lipingcoding thx for reporting the issue, I also observed some problems and still wondering what's wrong with pytorch version compared with tf implementation. Sorry to say but I still haven't figured it out yet. The only thing I can be sure currently is that original paper's hyperparameter setting could be be directly used for this codebase, as I fixed some leaky attention issue by using PyTorch's MHA, the parameter initialization issue still need to be elaborated but I haven't done it yet.

pmixer avatar Nov 13 '20 02:11 pmixer

btw @lipingcoding could u pls provide bit more information like log etc. to describe the observation that the model do not converge? Like uncomment https://github.com/pmixer/SASRec.pytorch/blob/30c43cf090d429480339ab18d43354b3e399bc29/main.py#L93 would print loss for iteration which should be informative.

pmixer avatar Nov 13 '20 03:11 pmixer

@lipingcoding try the newly updated code?

pmixer avatar Nov 13 '20 05:11 pmixer