Closed zenogantner closed 11 years ago
Rationale: fast sampling makes running BPR on smaller data sets faster -- but there it is not so much of a problem anyway. Also, we want to renew and generalize sampling with the new data model anyway, so we can remove this to simplify our code.
Done.
Rationale: fast sampling makes running BPR on smaller data sets faster -- but there it is not so much of a problem anyway. Also, we want to renew and generalize sampling with the new data model anyway, so we can remove this to simplify our code.