Introduction

This document describes how to run the Deep Learning Recommendation Model (DLRM) in the openEuler 22.03 LTS SP3 operating system (OS) running on the Kunpeng processor. The used deep learning framework is TensorFlow 2.13.0.

The DLRM is developed by Facebook to process large recommendation tasks. It combines deep learning and classical recommender system, and uses a hybrid input type (sparse feature and dense feature) method to process recommendation tasks. It is implemented using PyTorch. You need to use the tensorflow-dlrm open source repository as well to complete the porting.

For more information about the DLRM, visit https://github.com/facebookresearch/dlrm.

For more information about tensorflow-dlrm, visit https://github.com/NodLabs/tensorflow-dlrm.

Programming language: Python

Brief description: a deep learning recommendation model

Open source license: MIT License