NVIDIA-Merlin / Merlin

NVIDIA Merlin is an open source library providing end-to-end GPU-accelerated recommender systems, from feature engineering and preprocessing to training deep learning models and running inference in production.
Apache License 2.0
715 stars 111 forks source link

[QST] What is the best way of handling string UUIDs in Merlin? #1091

Closed BlakeB415 closed 6 months ago

BlakeB415 commented 6 months ago

Discussed in https://github.com/NVIDIA-Merlin/Merlin/discussions/1089

Originally posted by **BlakeB415** December 22, 2023 What would be the best way of going about this? I would rather not maintain a separate UUID to Int mapping table. My original idea was to either Hash or Categorify the UUID values for training, then have the ``user_id_raw`` and ``item_id_raw`` string values be returned from the Feast store during inference so that my application can get the raw UUIDs. However, Merlin does not support this paradigm as I get type errors during inference.
BlakeB415 commented 6 months ago

Hi, it's been a week. Does anyone know how to deal with this? Thank you.

rnyak commented 6 months ago

@BlakeB415 please share your error message and also a toy example for us to repro your issue.