search

massquantity / LibRecommender

Versatile End-to-End Recommender System
https://librecommender.readthedocs.io/
MIT License
336 stars 59 forks source link

lightgcn with bpr loss issue #464

Open rajanskumarsoni opened 3 months ago

rajanskumarsoni commented 3 months ago

When I ma running lightgcn model with negative and postive samples still it is saying 'negative_sample' parameter to be true. I want to make it false because I already have negative sample in my data.

massquantity commented 3 months ago

bpr loss belongs to pairwise loss, which is complicated to process with users providing both negative and positive samples. So it only supports negative sampling, see Caution in Loss.