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

1.85k stars 309 forks source link

Replace functools cache with lru_cache #967

Closed timmoon10 closed 3 months ago

timmoon10 commented 3 months ago

Description

functools.cache was added in Python 3.9, leading to problems with older Python versions (see https://github.com/NVIDIA/TransformerEngine/issues/958). The fix is to replace it with functools.lru_cache(maxsize=None), which is equivalent.

I've also made some minor stylistic changes in the affected code:

Alphabetized imports
Avoided importing individual functions, per Google's style guide

Type of change

[ ] Documentation change (change only to the documentation, either a fix or a new content)
[x] Bug fix (non-breaking change which fixes an issue)
[ ] New feature (non-breaking change which adds functionality)
[ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
[ ] Infra/Build change
[ ] Code refractor

Changes

Replaced functools.cache with functools.lru_cache(maxsize=None)

Checklist:

[x] I have read and followed the contributing guidelines
[x] The functionality is complete
[x] I have commented my code, particularly in hard-to-understand areas
[x] I have made corresponding changes to the documentation
[ ] My changes generate no new warnings
[ ] I have added tests that prove my fix is effective or that my feature works
[ ] New and existing unit tests pass locally with my changes

timmoon10 commented 3 months ago

/te-ci