PyTorch on Apple M1 MAX GPUs with SHARK – faster than TensorFlow-Metal

Re:PyTorch on Apple M1 MAX GPUs with SHARK – faster than Ten... by liquidmetal on Tuesday February 22, 2022 @10:54PM (#30437384)

How do your benchmarks compare against CoreML?

[ Parent | on HN ]

Re:PyTorch on Apple M1 MAX GPUs with SHARK – faster than Ten... by magic_at_nodai on Tuesday February 22, 2022 @11:48PM (#30437768)

I updated the blog with the reference. Basically it crashes to compile the model with https://github.com/NodLabs/shark-samples/blob/main/examples/.... The coremltools converter is very version specific (like all vendor conversion kits) and still on a version of TF I couldn't get on conda. Also it doesn't allow for training and only FP16 for inference with ANE. All our tests were with FP32.

//part of nod.ai/shark team.

[ Parent | on HN ]