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

When using Import Transformer_engine, many processes will be created

zte-tcb opened this issue · comments

commented

When I only import transformer_engine, it will create many processes. I found that this was pre created many compilation processes when using torch.compile. I want to know if these processes can stop after the first round of training.

TransformerEngine：1.4
torch: 2.1.0