[HN Gopher] PyTorch on Apple M1 Faster Than TensorFlow-Metal ___________________________________________________________________ PyTorch on Apple M1 Faster Than TensorFlow-Metal Author : yayr Score : 7 points Date : 2022-02-22 22:39 UTC (21 minutes ago) (HTM) web link (nod.ai) (TXT) w3m dump (nod.ai) | yayr wrote: | there is also some hint about performance/watt vs an NVIDIA A100 | using BERT models: | | > The A100 draws 131W peak during the training run, the M1MAX GPU | draws a maximum of 15.4 W. The A100 runs an iteration at 7ms vs | 104ms (with SHARK) and 196ms (with TF-Metal) ___________________________________________________________________ (page generated 2022-02-22 23:00 UTC)