Tuesday, October 4, 2022

Meta announces AITemplate, an open-source, PyTorch-based unified "inference" system for both AMD and Nvidia GPU hardware, helping code run 4x-12x faster (Reuters)

Reuters:
Meta announces AITemplate, an open-source, PyTorch-based unified “inference” system for both AMD and Nvidia GPU hardware, helping code run 4x-12x faster  —  Facebook parent Meta Platforms Inc (META.O)said on Monday it has launched a new set of free software tools …



at
Labels:

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)

SpaceX's S-1 excerpts list "manufacturing our own GPUs" among the "substantial capital expenditures" it is undertaking, with the size of the expenditure TBD (Reuters)

Reuters : SpaceX's S-1 excerpts list “manufacturing our own GPUs” among the “substantial capital expenditures” it is undertaking, wit...