Explore using conda as a package/dependency manager and avenue to distribute TTMetal

tenstorrent / tt-metal

:metal: TT-NN operator library, and TT-Metalium low level kernel programming model.

Apache License 2.0

397 stars 49 forks source link

Explore using conda as a package/dependency manager and avenue to distribute TTMetal #2879

Open TT-billteng opened 11 months ago

TT-billteng commented 11 months ago

Conda's tools and ecosystem enable multi-arch/OS deployment and is quite popular in the ML space. Pytorch, Tensorflow, Intel, Nvidia all have custom channels to control their dependencies, and distribute their software to the public.

tt-rkim commented 6 months ago

Bringing this up again, as we're talking about supporting multiple Ubuntu and Python versions