2 ms·
ONNX Runtime purports to support CoreML: https://onnxruntime.ai/docs/execution-providers/CoreML-ExecutionProvider.html https://onnxruntime.ai/docs/execution-pro
by daemonologist 1y ago
ONNX Runtime purports to support CoreML: https://onnxruntime.ai/docs/execution-providers/CoreML-ExecutionProvider.html https://onnxruntime.ai/docs/execution-providers/CoreML-Execu... , which gives a decent amount of compatibility for inference. I have no idea to what extent workloads actually end up on the ANE though.
(Unfortunately ONNX doesn't support Vulkan, which limits it on other platforms. It's always something...)