SparseOperationKit (SOK) is a python package wrapped GPU accelerated operations dedicated for sparse training / inference cases.
Project description
SparseOperationKit (SOK) is a python package wrapped GPU accelerated operations dedicated for sparse training / inference cases. It is designed to be compatible with common DeepLearning (DL) frameworks, for instance, TensorFlow. Most of the algorithm implementations in SOK are extracted from HugeCTR, which is a GPU-accelerated recommender framework designed to distribute training across multiple GPUs and nodes and estimate Click-Through Rates (CTRs). If you are looking for a very efficient solution for CTRs, please check HugeCTR.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.