Giters

NVIDIA / TransformerEngine

A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs, to provide better performance with lower memory utilization in both training and inference.

Home Page:https://docs.nvidia.com/deeplearning/transformer-engine/user-guide/index.html

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Can TE optimize the find cudnn?

MARD1NO opened this issue · comments

commented

Maybe user do not install cuda/cudnn toolkit independently, but pytorch will pack their own cudnn in installation package.

I think TE can set the default cudnn path by pytorch install dir, otherwise user need to specify the cudnn include directory by setting ENV_VAR CUDNN_INCLUDE_DIR.

commented

Hi @MARD1NO, could you please give #831 a try? Thanks!