gitmyhub

cuBERT

C++ ★ 547 updated 5y ago

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

No plain-English explanation yet — one is being written right now. Check back in a minute.