MLCommons Group Aims to Benchmark AI Performance on Consumer Devices
-
MLCommons is launching a new working group, MLPerf Client, to establish AI benchmarks for laptops, desktops, and workstations.
-
The first benchmark will focus on text generation models like Meta's Llama 2, which is already used in MLCommons' datacenter benchmarks.
-
The goal is to make it easier to compare consumer PCs on their ability to run AI apps and workloads.
-
Members include AMD, Arm, Dell, Intel, Nvidia, Qualcomm, but not Apple which is not part of MLCommons.
-
The benchmarks could help buyers evaluate devices for expected AI performance, like PC build comparison tools do for gaming and other workflows.