Dear users, we have installed a new software: flash-attention 2.6.3-CUDA-12.1.1
:
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
flash-attention: flash-attention/2.6.3-CUDA-12.1.1
---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Description:
Fast and memory-efficient exact attention.
You will need to load all module(s) on any one of the lines below before the "flash-attention/2.6.3-CUDA-12.1.1" module is available to load.
GCC/12.3.0 OpenMPI/4.1.5
Help:
Description
===========
Fast and memory-efficient exact attention.
More information
================
- Homepage: https://github.com/Dao-AILab/flash-attention
Included extensions
===================
flash-attention-2.6.3
Best,
HPC team