/cuBERT

Fast implementation of BERT inference directly on NVIDIA (CUDA, CUBLAS) and Intel MKL

Primary LanguageC++MIT LicenseMIT

No issues in this repository yet.