You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In our offline evaluation (precision, recall), we currently iterate through all users and run get_recommendations for each user. This is slow.
Instead we should generate the full user x item recommendation set in one matrix multiplication operation. This involves changing get_precision_recall in https://github.com/blockchain-etl/token-recommender/blob/master/wals_ml_engine/trainer/wals.py
The text was updated successfully, but these errors were encountered:
In our offline evaluation (precision, recall), we currently iterate through all users and run
get_recommendations
for each user. This is slow.Instead we should generate the full user x item recommendation set in one matrix multiplication operation. This involves changing
get_precision_recall
in https://github.com/blockchain-etl/token-recommender/blob/master/wals_ml_engine/trainer/wals.pyThe text was updated successfully, but these errors were encountered: