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.