При попытке обучения с API обнаружения объектов Tensorflow через Google Colab я получил эту ошибку. Я использую TensorFlow==1.5, и я уже добавил следующие пути (модели, исследования и каталоги slim):
%set_env PYTHONPATH=/content/gdrive/My Drive/Computer_vision_with_deep_learning/TFOD/models:/content/gdrive/My Drive/Computer_vision_with_deep_learning/TFOD/models/research:/content/gdrive/My Drive/Computer_vision_with_deep_learning/TFOD/models/research/slim
Во время выполнения команды обучения (!python training/train.py --logtostderr --train_dir=training/ --pipeline_config_path=training/ssd_mobilenet_v1_coco.config) я получил следующее подробное описание ошибки :
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:493: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint8 = np.dtype([("qint8", np.int8, 1)])
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:494: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint8 = np.dtype([("quint8", np.uint8, 1)])
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:495: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint16 = np.dtype([("qint16", np.int16, 1)])
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:496: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_quint16 = np.dtype([("quint16", np.uint16, 1)])
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:497: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
_np_qint32 = np.dtype([("qint32", np.int32, 1)])
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/dtypes.py:502: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.
np_resource = np.dtype([("resource", np.ubyte, 1)])
Traceback (most recent call last):
File "training/train.py", line 26, in <module>
from object_detection import model_lib
File "/content/gdrive/My Drive/Computer_vision_with_deep_learning/TFOD/training/object_detection/model_lib.py", line 28, in <module>
from object_detection import exporter as exporter_lib
File "/content/gdrive/My Drive/Computer_vision_with_deep_learning/TFOD/training/object_detection/exporter.py", line 23, in <module>
from object_detection.builders import model_builder
File "/content/gdrive/My Drive/Computer_vision_with_deep_learning/TFOD/training/object_detection/builders/model_builder.py", line 47, in <module>
from object_detection.models import center_net_hourglass_feature_extractor
ImportError: cannot import name 'center_net_hourglass_feature_extractor'
Как исправить?