search

microsoft / BitBLAS

BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.
MIT License
359 stars 29 forks source link

[BitNet] Disable accelerate for BitNET #36

Closed LeiWang1999 closed 4 months ago

LeiWang1999 commented 4 months ago
  1. add the requirement.txt for bitnet integration
  2. disable accelerate as it might enable cuda graph.