installation failed due to demand of old flash attention

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.

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

Apache License 2.0

1.6k stars 255 forks source link

installation failed due to demand of old flash attention #928

Closed saurabh-kataria closed 2 weeks ago

saurabh-kataria commented 2 weeks ago

Installation fails due to requirement of older flash-attn!=2.0.9,!=2.1.0,<=2.4.2