recommenders-team/recommenders

[ASK] Azure ML Compute based on algo and users / item size

Aperta

#896 aperta il 7 ago 2019

 (2 commenti) (0 reazioni) (0 assegnatari)Python (3319 fork)batch import
help wanted

Metriche repository

Star
 (21.866 stelle)
Metriche merge PR
 (Metriche PR in attesa)

Descrizione

Description

I'm trying to run RBM on a problem with about 20K users and 40K items. Is there a way to quickly figure out what type of Azure Machine Learning Compute I would need to handle this size of problem, especially for the GPU. I can run this on an Standard NC12 on the CPU, but when I try to run it with the GPU I get an OOM error or the run fails with the last log entry somewhere deep in Tensorflow.

Other Comments

Guida contributor