运行TFLite Model Maker代码时遇array_record_module导入错误

运行TensorFlow Lite Model Maker官方图像分类示例代码时触发ImportError，错误提示array_record_module.so不是mach-o格式文件（macOS系统环境）。

运行的代码 #

from tflite_model_maker import image_classifier
from tflite_model_maker.image_classifier import DataLoader

# 加载设备端ML应用的输入数据
data = DataLoader.from_folder('flower_photos/')
train_data, test_data = data.split(0.9)

# 定制TensorFlow模型
model = image_classifier.create(train_data)

# 评估模型
loss, accuracy = model.evaluate(test_data)

# 导出TensorFlow Lite模型和标签文件到指定目录
model.export(export_dir='/tmp/')

错误日志 #

2023-05-14 13:43:07.639151: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
/usr/local/lib/python3.9/site-packages/tensorflow_addons/utils/tfa_eol_msg.py:23: UserWarning: 

TensorFlow Addons (TFA) has ended development and introduction of new features.
TFA has entered a minimal maintenance and release mode until a planned end of life in May 2024.
Please modify downstream libraries to take dependencies from other repositories in our TensorFlow community (e.g. Keras, Keras-CV, and Keras-NLP). 

For more information see: https://github.com/tensorflow/addons/issues/2807 

  warnings.warn(
Traceback (most recent call last):
  File "/Users/apple/Documents/modelmakersrc/maketflitemodel.py", line 3, in <module>
    from tflite_model_maker import model_spec
  File "/usr/local/lib/python3.9/site-packages/tflite_model_maker/__init__.py", line 44, in <module>
    from tflite_model_maker import audio_classifier
  File "/usr/local/lib/python3.9/site-packages/tflite_model_maker/audio_classifier/__init__.py", line 24, in <module>
    from tensorflow_examples.lite.model_maker.core.data_util.audio_dataloader import DataLoader
  File "/usr/local/lib/python3.9/site-packages/tensorflow_examples/lite/model_maker/core/data_util/audio_dataloader.py", line 27, in <module>
    from tensorflow_examples.lite.model_maker.core.task.model_spec import audio_spec
  File "/usr/local/lib/python3.9/site-packages/tensorflow_examples/lite/model_maker/core/task/model_spec/__init__.py", line 24, in <module>
    from tensorflow_examples.lite.model_maker.core.task.model_spec import text_spec
  File "/usr/local/lib/python3.9/site-packages/tensorflow_examples/lite/model_maker/core/task/model_spec/text_spec.py", line 44, in <module>
    from official.nlp.data import classifier_data_lib
  File "/usr/local/lib/python3.9/site-packages/official/nlp/data/classifier_data_lib.py", line 28, in <module>
    import tensorflow_datasets as tfds
  File "/usr/local/lib/python3.9/site-packages/tensorflow_datasets/__init__.py", line 43, in <module>
    import tensorflow_datasets.core.logging as _tfds_logging
  File "/usr/local/lib/python3.9/site-packages/tensorflow_datasets/core/__init__.py", line 22, in <module>
    from tensorflow_datasets.core import community
  File "/usr/local/lib/python3.9/site-packages/tensorflow_datasets/core/community/__init__.py", line 18, in <module>
    from tensorflow_datasets.core.community.huggingface_wrapper import mock_builtin_to_use_gfile
  File "/usr/local/lib/python3.9/site-packages/tensorflow_datasets/core/community/huggingface_wrapper.py", line 31, in <module>
    from tensorflow_datasets.core import dataset_builder
  File "/usr/local/lib/python3.9/site-packages/tensorflow_datasets/core/dataset_builder.py", line 34, in <module>
    from tensorflow_datasets.core import dataset_info
  File "/usr/local/lib/python3.9/site-packages/tensorflow_datasets/core/dataset_info.py", line 47, in <module>
    from tensorflow_datasets.core import file_adapters
  File "/usr/local/lib/python3.9/site-packages/tensorflow_datasets/core/file_adapters.py", line 29, in <module>
    from array_record.python import array_record_module
ImportError: dlopen(/usr/local/lib/python3.9/site-packages/array_record/python/array_record_module.so, 0x0002): tried: '/usr/local/lib/python3.9/site-packages/array_record/python/array_record_module.so' (not a mach-o file)

关键依赖包版本 #

• array-record: 0.2.0
• tensorflow: 2.12.0
• tensorflow-datasets: 4.9.1
• tflite-support: 0.1.0a1

错误根源是array-record包安装了Linux平台的二进制版本（.so格式），而macOS需要mach-o格式的文件。按以下步骤修复：

1. 卸载不兼容的array-record包

pip uninstall -y array-record

2. 安装适配macOS的版本
   降级到兼容的版本，或安装支持macOS架构的包：

pip install array-record==0.1.1

若上述命令无效，尝试用conda安装（需激活conda环境）：

conda install -c conda-forge array-record

3. 验证安装有效性
   运行以下代码确认无导入错误：

from array_record.python import array_record_module

4. 可选：重新安装tensorflow-datasets
   若问题仍存在，重新安装tensorflow-datasets确保依赖匹配：

pip uninstall -y tensorflow-datasets
pip install tensorflow-datasets==4.9.1

内容的提问来源于stack exchange，提问作者Charwaka