Crash when loading tensorflow

[wangjia184]

• Python version : 3.9.12
• Platform: macOS-12.4-arm64-arm-64bit
• Tensorflow version: 2.9.2
• PyO3 version: 0.17.1

The python code works well. But when I try to load the module from file in PyO3, it crashes.

    Python::with_gil(|py| {
        let module = PyModule::from_code(py, &source_code, "", "")?;
        let train_fn: Py<PyAny> = module.getattr("train")?.into();

        Ok(())
    })

Metal device set to: Apple M1

systemMemory: 16.00 GB
maxCacheSize: 5.33 GB

2022-09-08 10:23:57.897901: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:305] Could not identify NUMA node of platform GPU ID 0, defaulting to 0. Your kernel may not have been built with NUMA support.
2022-09-08 10:23:57.898139: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:271] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 0 MB memory) -> physical PluggableDevice (device: 0, name: METAL, pci bus id: <undefined>)
2022-09-08 10:23:58.043489: E tensorflow/core/lib/monitoring/collection_registry.cc:81] Cannot register 2 metrics with the same name: /tensorflow/core/op_expansion/node_counter
2022-09-08 10:23:58.043522: E tensorflow/core/lib/monitoring/collection_registry.cc:81] Cannot register 2 metrics with the same name: /tensorflow/core/op_expansion/graph_counter
2022-09-08 10:23:58.043531: E tensorflow/core/lib/monitoring/collection_registry.cc:81] Cannot register 2 metrics with the same name: /tensorflow/core/op_expansion/op_counter
2022-09-08 10:23:58.044998: E tensorflow/core/lib/monitoring/collection_registry.cc:81] Cannot register 2 metrics with the same name: /tensorflow/core/xla_launch_counter
2022-09-08 10:23:58.045166: F tensorflow/core/framework/op.cc:63] Non-OK-status: RegisterAlreadyLocked(op_data_factory) status: ALREADY_EXISTS: Op with name XlaLaunch

Seems it crashed in the reason of

RegisterAlreadyLocked(op_data_factory) status: ALREADY_EXISTS: Op with name XlaLaunch

Does someone have an hint how to solve this error?