我一直在尝试TensorFlow lite，并且在Android上也遇到了问题，因此我尝试测试.pb和.tflite模型，以了解与tflite_diff_example_test是否存在差异
我已经重新训练了我转换为.tflite的mobilenet_v1_100_224
我在以下MacOS 10.13.3上运行：
bazel build tensorflow/contrib/lite/testing/tflite_diff_example_test.cc
bazel-bin/tensorflow/contrib/lite/testing/tflite_diff_example_test --tensorflow_model=../new_training_dir/retrainedMNetV1_6_8000steps.pb --tflite_model=../new_training_dir/retrainedMNetV1_6_8000steps.tflite
我得到以下输出：
I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.2 AVX
Failed to invoke interpreter
Failed to invoke interpreter
...
Failed to invoke interpreter
Num errors: 100
我不知道问题出在哪里