Coordinated software tools could cut AI training energy use by up to 50%

T2Covered by 1 source · first reported by Phys.org / ScienceDaily (tech) at 12 Oct, 20:20 UTC

Around half of the energy used to train AI can be saved without affecting speed or performance, using software developed at the University of Michigan Engineering. Now, U-M researchers are working with NVIDIA to integrate their new approach into Megatron-LM, NVIDIA's open-source platform for…