Following this video "Accelerate machine learning with Metal" & this sample code "Customizing a TensorFlow operation" My Environment: macOS 13.1 Python 3.8.15 tensorflow-deps 2.9.0 tensorflow-macos 2.9.0 tensorflow-metal 0.5.1 after running make at hash_encoder, the hash_encoder_kernel.so appears. Then I tried to run python tiny_nerf_hash.py but it didn't work. Unlike the python tiny_nerf_mlp.py, it works normally but it didn't take the advantage of accelerating.

I just install the latest version of TensorFlow and TensorFlow-metal on my Mac (M1 Max macOS 12.5) and try to run the demo code of "Customizing A TensorFlow Operation". But it reports the error like this: $ python tiny_nerf_hash.py Checking whether it needs to recompile... Loading the .so binary... Traceback (most recent call last): File "/Users/-/Downloads/CustomizingATensorFlowOperation/tiny_nerf_hash.py", line 22, in <module> from hash_encoder import HashEncoder File "/Users/-/Downloads/CustomizingATensorFlowOperation/./hash_encoder/hash_encoder.py", line 37, in <module> _backend = tf.load_op_library(os.path.join(pathlib.Path( File "/Users/-/miniforge3/envs/instsha/lib/python3.9/site-packages/tensorflow/python/framework/load_library.py", line 54, in load_op_library lib_handle = py_tf.TF_LoadLibrary(library_filename) tensorflow.python.framework.errors_impl.NotFoundError: dlopen(/Users/-/Downloads/CustomizingATensorFlowOperation/hash_encoder/hash_encoder_kernel.so, 0x0006): symbol not found in flat namespace (__ZN10tensorflow14kernel_factory17OpKernelRegistrar12InitInternalEPKNS_9KernelDefEN4absl12lts_2022062311string_viewENSt3__110unique_ptrINS0_15OpKernelFactoryENS8_14default_deleteISA_EEEE)

Silly question but does anyone know where can I find the code for the demos in this video? https://developer.apple.com/videos/play/wwdc2022/10063/ I am trying to replicate the distributed training but running into version errors with Horovod.