Need of pre-trained model to save time and reproduce experiments

[tommasocarraro]

Dear collegues, I need help. I'm developing a new method for context-aware recommendation and I need to reproduce your experiments as fast as possible to compare my method with yours. I need this because it seems your method is one the best in the literature. I tried to run the code but I needed to convert everything to tf 2.0 and then the training process stucked on the validation after 10 epochs. Ithink that to be able to reproduce your experiments I need your exact model. Please help a researcher, thank you. If it is possible for you to provide the preprocessed movielens and also the best hyperparameters for every dataset it will be perfect.

[tommasocarraro]

Hi guys, do you have at least a way to perform the evaluation faster? Even in GPU it takes hours and hours to evaluate because you have 12000 users leave-one-out evaluation.