如何在tensorflow 1.13.1和Python3.7中解决“导入pycocotools._mask不是有效的win32应用程序”？

Question

我正在使用object detection raccoon by datitran(https://github.com/datitran/raccoon_dataset)来训练我自己的数据集，但遇到了一个问题。

环境信息：

操作系统:Windows 10 64位

CUDA和CUDNN: CUDA 10/CUDNN 7.5.0

Tensorflow版本:1.13.1(GPU)

Python版本:3.7.0 64位(win32上的Anaconda，Inc. )

Model:ssd_mobilenet_v1_coco

问题：

我想开始训练

python model_main.py 
--pipeline_config_path=Z:/Train/ssd_mobilenet_v1_coco_ship.config 
--model_dir=Z:/Train/train 
--alsologtostderr

遇到了一个错误

    Traceback (most recent call last):
  File "model_main.py", line 26, in <module>
    from object_detection import model_lib
  File "D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\model_lib.py", line 27, in <module>
    from object_detection import eval_util
  File "D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\eval_util.py", line 27, in <module>
    from object_detection.metrics import coco_evaluation
  File "D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\metrics\coco_evaluation.py", line 20, in <module>
    from object_detection.metrics import coco_tools
  File "D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\metrics\coco_tools.py", line 47, in <module>
    from pycocotools import coco
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\pycocotools\coco.py", line 56, in <module>
    from . import mask as maskUtils
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\pycocotools\mask.py", line 3, in <module>
    import pycocotools._mask as _mask
ImportError: DLL load failed: %1 is not a valid Win32 application.

此外，我还在~\research\object_detection\legacy中尝试了旧版本的命令train.py

python train.py 
--logtostderr 
--pipeline_config_path=Z:/Train/ssd_mobilenet_v1_coco_ship.config  
--train_dir=Z:/Train/train

但它也不起作用：

WARNING: The TensorFlow contrib module will not be included in TensorFlow 2.0.
For more information, please see:
  * https://github.com/tensorflow/community/blob/master/rfcs/20180907-contrib-sunset.md
  * https://github.com/tensorflow/addons
If you depend on functionality not listed there, please file an issue.

WARNING:tensorflow:From D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\platform\app.py:125: main (from __main__) is deprecated and will be removed in a future version.
Instructions for updating:
Use object_detection/model_main.py.
WARNING:tensorflow:From D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\legacy\trainer.py:266: create_global_step (from tensorflow.contrib.framework.python.ops.variables) is deprecated and will be removed in a future version.
Instructions for updating:
Please switch to tf.train.create_global_step
WARNING:tensorflow:From D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.
Instructions for updating:
Colocations handled automatically by placer.
WARNING:tensorflow:num_readers has been reduced to 1 to match input file shards.
WARNING:tensorflow:From D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\builders\dataset_builder.py:80: parallel_interleave (from tensorflow.contrib.data.python.ops.interleave_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use `tf.data.experimental.parallel_interleave(...)`.
Traceback (most recent call last):
  File "train.py", line 186, in <module>
    tf.app.run()
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\platform\app.py", line 125, in run
    _sys.exit(main(argv))
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\util\deprecation.py", line 324, in new_func
    return func(*args, **kwargs)
  File "train.py", line 182, in main
    graph_hook_fn=graph_rewriter_fn)
  File "D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\legacy\trainer.py", line 280, in train
    train_config.prefetch_queue_capacity, data_augmentation_options)
  File "D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\legacy\trainer.py", line 59, in create_input_queue
    tensor_dict = create_tensor_dict_fn()
  File "train.py", line 123, in get_next
    dataset_builder.build(config)).get_next()
  File "D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\builders\dataset_builder.py", line 134, in build
    config.input_path[:], input_reader_config)
  File "D:\Anaconda\envs\tensorflow\Lib\models\research\object_detection\builders\dataset_builder.py", line 80, in read_dataset
    sloppy=config.shuffle))
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 1605, in apply
    return DatasetV1Adapter(super(DatasetV1, self).apply(transformation_func))
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 1127, in apply
    dataset = transformation_func(self)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\experimental\ops\interleave_ops.py", line 88, in _apply_fn
    buffer_output_elements, prefetch_input_elements)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\readers.py", line 133, in __init__
    cycle_length, block_length)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 2827, in __init__
    super(InterleaveDataset, self).__init__(input_dataset, map_func)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 2798, in __init__
    map_func, self._transformation_name(), dataset=input_dataset)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 2124, in __init__
    self._function.add_to_graph(ops.get_default_graph())
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\function.py", line 490, in add_to_graph
    self._create_definition_if_needed()
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\function.py", line 341, in _create_definition_if_needed
    self._create_definition_if_needed_impl()
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\function.py", line 355, in _create_definition_if_needed_impl
    whitelisted_stateful_ops=self._whitelisted_stateful_ops)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\function.py", line 883, in func_graph_from_py_func
    outputs = func(*func_graph.inputs)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 2099, in tf_data_structured_function_wrapper
    ret = func(*nested_args)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\readers.py", line 247, in __init__
    filenames, compression_type, buffer_size, num_parallel_reads)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\readers.py", line 212, in __init__
    self._impl = filenames.flat_map(read_one_file)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 1005, in flat_map
    return FlatMapDataset(self, map_func)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 2798, in __init__
    map_func, self._transformation_name(), dataset=input_dataset)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 2124, in __init__
    self._function.add_to_graph(ops.get_default_graph())
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\function.py", line 490, in add_to_graph
    self._create_definition_if_needed()
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\function.py", line 341, in _create_definition_if_needed
    self._create_definition_if_needed_impl()
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\function.py", line 355, in _create_definition_if_needed_impl
    whitelisted_stateful_ops=self._whitelisted_stateful_ops)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\function.py", line 883, in func_graph_from_py_func
    outputs = func(*func_graph.inputs)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\dataset_ops.py", line 2099, in tf_data_structured_function_wrapper
    ret = func(*nested_args)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\readers.py", line 209, in read_one_file
    return _TFRecordDataset(filename, compression_type, buffer_size)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\ops\readers.py", line 111, in __init__
    argument_dtype=dtypes.string)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\data\util\convert.py", line 35, in optional_param_to_tensor
    argument_default, dtype=argument_dtype, name=argument_name)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\constant_op.py", line 245, in constant
    allow_broadcast=True)
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\constant_op.py", line 283, in _constant_impl
    allow_broadcast=allow_broadcast))
  File "D:\Anaconda\envs\tensorflow\lib\site-packages\tensorflow\python\framework\tensor_util.py", line 501, in make_tensor_proto
    (dtype, nparray.dtype, values))
TypeError: Incompatible types: <dtype: 'string'> vs. object. Value is

我应该怎么做才能让它们中的任何一个工作？

Answer 1

你需要安装适用于pycocotools的VİSUAL C++ 2015，但是当我安装visual c++时，它不能工作。然后我编辑pycocotools以使用等待片刻im加载它github

编辑:已加载的https://github.com/rifatalptekincetin/pycocotools-edited

下载pycocotools文件夹，将其替换为research/object_detection文件路径必须为：

models-master/resarch/object_detection/pycocotools/__init__.py #etc.

Answer 2

至少有一个Python3和pycocotools之间的不兼容也在GitHub上生成了一个拉取请求。其中之一是：

​

​

有关更多信息，请访问：

Pull Request for pycocotools library

Answer 3

嘿，对于任何使用windows的人来说，这真的起作用了。

1)使用默认选择从https://go.microsoft.com/fwlink/?LinkId=691126安装Visual C++ 2015生成工具。

2)转到C:\Program Files (x86)\Microsoft C++ Build Tools并运行vcbuildtools_msbuild.bat

3)然后在Anaconda中复制并运行此命令(您必须已安装git ...如果不是只使用pip install git)

pip安装

git+https://github.com/philferriere/cocoapi.git#egg=pycocotools^&subdirectory=PythonAPI

希望它也适用于您:)