Question

我正在尝试在我自己的数据集上训练tensorflow对象检测。

我做了什么？

使用ssd_mobilenet_v1_pets.config作为基础来创建自己的管道配置。修改了num_classes和所有其他路径特定部分以匹配我的环境。
来自tensorflow model zoo的ssd_mobilenet_v1_coco作为检查点
使用所有标签创建标签地图文件（第一个索引从1开始）
从我的数据集创建了一个TFRecord文件（该脚本基于tensorflow sample script）

出了什么问题？

开始培训时： python tensorflow_models/research/object_detection/train.py --pipeline_config_path=/home/playground/ssd_mobilenet_v1.config --train_dir=/tmp/bla/我得到以下回溯：

Traceback (most recent call last):
  File "tensorflow_models/research/object_detection/train.py", line 198, in <module>
    tf.app.run()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/app.py", line 48, in run
    _sys.exit(main(_sys.argv[:1] + flags_passthrough))
  File "tensorflow_models/research/object_detection/train.py", line 194, in main
    worker_job_name, is_chief, FLAGS.train_dir)
  File "/home/playground/tensorflow_models/research/object_detection/trainer.py", line 296, in train
    saver=saver)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/slim/python/slim/learning.py", line 767, in train
    sv.stop(threads, close_summary_writer=True)
  File "/usr/lib/python2.7/contextlib.py", line 35, in __exit__
    self.gen.throw(type, value, traceback)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/supervisor.py", line 964, in managed_session
    self.stop(close_summary_writer=close_summary_writer)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/supervisor.py", line 792, in stop
    stop_grace_period_secs=self._stop_grace_secs)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/coordinator.py", line 389, in join
    six.reraise(*self._exc_info_to_raise)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/queue_runner_impl.py", line 238, in _run
    enqueue_callable()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 1235, in _single_operation_run
    target_list_as_strings, status, None)
  File "/usr/lib/python2.7/contextlib.py", line 24, in __exit__
    self.gen.next()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/errors_impl.py", line 466, in raise_exception_on_not_ok_status
    pywrap_tensorflow.TF_GetCode(status))
tensorflow.python.framework.errors_impl.InvalidArgumentError: indices[0] = 2 is not in [0, 1)
         [[Node: cond/RandomCropImage/PruneCompleteleyOutsideWindow/Gather/Gather_1 = Gather[Tindices=DT_INT64, Tparams=DT_INT64, validate_indices=true, _device="/job:localhost/replica:0/task:0/cpu:0"](cond/RandomCropImage/PruneCompleteleyOutsideWindow/Gather/Gather_1/Switch:1, cond/RandomCropImage/PruneCompleteleyOutsideWindow/Reshape)]]

不幸的是，我不知道张量流是什么想告诉我这个追溯，也不知道我应该从哪里开始寻找我的错误。我已经检查了每一步可能出现的错误，但到目前为止还找不到任何错误。

编辑：我也尝试使用this配置文件，正如@eshirima所提议的那样。我再次更改了num_classes参数以及标有PATH_TO_BE_CONFIGURED的所有其他参数。但是，现在它失败并显示以下错误消息：

INFO:tensorflow:Starting Queues.
INFO:tensorflow:global_step/sec: 0
INFO:tensorflow:Error reported to Coordinator: <class 'tensorflow.python.framework.errors_impl.InvalidArgumentError'>, indices[0] = 2 is not in [0, 1)
         [[Node: Loss/Gather_29 = Gather[Tindices=DT_INT32, Tparams=DT_FLOAT, validate_indices=true, _device="/job:localhost/replica:0/task:0/cpu:0"](Loss/Pad_5, Loss/Reshape_47)]]

Caused by op u'Loss/Gather_29', defined at:
  File "tensorflow_models/research/object_detection/train.py", line 198, in <module>
    tf.app.run()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/app.py", line 48, in run
    _sys.exit(main(_sys.argv[:1] + flags_passthrough))
  File "tensorflow_models/research/object_detection/train.py", line 194, in main
    worker_job_name, is_chief, FLAGS.train_dir)
  File "/home/playground/tensorflow_models/research/object_detection/trainer.py", line 192, in train
    clones = model_deploy.create_clones(deploy_config, model_fn, [input_queue])
  File "/home/playground/tensorflow_models/research/slim/deployment/model_deploy.py", line 193, in create_clones
    outputs = model_fn(*args, **kwargs)
  File "/home/playground/tensorflow_models/research/object_detection/trainer.py", line 133, in _create_losses
    losses_dict = detection_model.loss(prediction_dict)
  File "/home/playground/tensorflow_models/research/object_detection/meta_architectures/ssd_meta_arch.py", line 411, in loss
    self.groundtruth_lists(fields.BoxListFields.classes))
  File "/home/playground/tensorflow_models/research/object_detection/meta_architectures/ssd_meta_arch.py", line 485, in _assign_targets
    groundtruth_classes_with_background_list)
  File "/home/playground/tensorflow_models/research/object_detection/core/target_assigner.py", line 438, in batch_assign_targets
    anchors, gt_boxes, gt_class_targets)
  File "/home/playground/tensorflow_models/research/object_detection/core/target_assigner.py", line 154, in assign
    match)
  File "/home/playground/tensorflow_models/research/object_detection/core/target_assigner.py", line 250, in _create_classification_targets
    matched_cls_targets = tf.gather(groundtruth_labels, matched_gt_indices)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/array_ops.py", line 2409, in gather
    validate_indices=validate_indices, name=name)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/gen_array_ops.py", line 1219, in gather
    validate_indices=validate_indices, name=name)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/op_def_library.py", line 767, in apply_op
    op_def=op_def)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 2630, in create_op
    original_op=self._default_original_op, op_def=op_def)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 1204, in __init__
    self._traceback = self._graph._extract_stack()  # pylint: disable=protected-access

InvalidArgumentError (see above for traceback): indices[0] = 2 is not in [0, 1)
         [[Node: Loss/Gather_29 = Gather[Tindices=DT_INT32, Tparams=DT_FLOAT, validate_indices=true, _device="/job:localhost/replica:0/task:0/cpu:0"](Loss/Pad_5, Loss/Reshape_47)]]

Traceback (most recent call last):
  File "tensorflow_models/research/object_detection/train.py", line 198, in <module>
    tf.app.run()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/app.py", line 48, in run
    _sys.exit(main(_sys.argv[:1] + flags_passthrough))
  File "tensorflow_models/research/object_detection/train.py", line 194, in main
    worker_job_name, is_chief, FLAGS.train_dir)
  File "/home/playground/tensorflow_models/research/object_detection/trainer.py", line 296, in train
    saver=saver)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/contrib/slim/python/slim/learning.py", line 767, in train
    sv.stop(threads, close_summary_writer=True)
  File "/usr/lib/python2.7/contextlib.py", line 35, in __exit__
    self.gen.throw(type, value, traceback)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/supervisor.py", line 964, in managed_session
    self.stop(close_summary_writer=close_summary_writer)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/supervisor.py", line 792, in stop
    stop_grace_period_secs=self._stop_grace_secs)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/coordinator.py", line 389, in join
    six.reraise(*self._exc_info_to_raise)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/coordinator.py", line 296, in stop_on_exception
    yield
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/coordinator.py", line 494, in run
    self.run_loop()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/training/supervisor.py", line 994, in run_loop
    self._sv.global_step])
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 895, in run
    run_metadata_ptr)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 1124, in _run
    feed_dict_tensor, options, run_metadata)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 1321, in _do_run
    options, run_metadata)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 1340, in _do_call
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: indices[0] = 2 is not in [0, 1)
         [[Node: Loss/Gather_29 = Gather[Tindices=DT_INT32, Tparams=DT_FLOAT, validate_indices=true, _device="/job:localhost/replica:0/task:0/cpu:0"](Loss/Pad_5, Loss/Reshape_47)]]

Caused by op u'Loss/Gather_29', defined at:
  File "tensorflow_models/research/object_detection/train.py", line 198, in <module>
    tf.app.run()
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/platform/app.py", line 48, in run
    _sys.exit(main(_sys.argv[:1] + flags_passthrough))
  File "tensorflow_models/research/object_detection/train.py", line 194, in main
    worker_job_name, is_chief, FLAGS.train_dir)
  File "/home/playground/tensorflow_models/research/object_detection/trainer.py", line 192, in train
    clones = model_deploy.create_clones(deploy_config, model_fn, [input_queue])
  File "/home/playground/tensorflow_models/research/slim/deployment/model_deploy.py", line 193, in create_clones
    outputs = model_fn(*args, **kwargs)
  File "/home/playground/tensorflow_models/research/object_detection/trainer.py", line 133, in _create_losses
    losses_dict = detection_model.loss(prediction_dict)
  File "/home/playground/tensorflow_models/research/object_detection/meta_architectures/ssd_meta_arch.py", line 411, in loss
    self.groundtruth_lists(fields.BoxListFields.classes))
  File "/home/playground/tensorflow_models/research/object_detection/meta_architectures/ssd_meta_arch.py", line 485, in _assign_targets
    groundtruth_classes_with_background_list)
  File "/home/playground/tensorflow_models/research/object_detection/core/target_assigner.py", line 438, in batch_assign_targets
    anchors, gt_boxes, gt_class_targets)
  File "/home/playground/tensorflow_models/research/object_detection/core/target_assigner.py", line 154, in assign
    match)
  File "/home/playground/tensorflow_models/research/object_detection/core/target_assigner.py", line 250, in _create_classification_targets
    matched_cls_targets = tf.gather(groundtruth_labels, matched_gt_indices)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/array_ops.py", line 2409, in gather
    validate_indices=validate_indices, name=name)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/ops/gen_array_ops.py", line 1219, in gather
    validate_indices=validate_indices, name=name)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/op_def_library.py", line 767, in apply_op
    op_def=op_def)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 2630, in create_op
    original_op=self._default_original_op, op_def=op_def)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 1204, in __init__
    self._traceback = self._graph._extract_stack()  # pylint: disable=protected-access

InvalidArgumentError (see above for traceback): indices[0] = 2 is not in [0, 1)
         [[Node: Loss/Gather_29 = Gather[Tindices=DT_INT32, Tparams=DT_FLOAT, validate_indices=true, _device="/job:localhost/replica:0/task:0/cpu:0"](Loss/Pad_5, Loss/Reshape_47)]]

编辑添加了一些代码，显示了TFRecord文件的生成方式。整个脚本有点长，但我试图将其剪切为仅显示相关部分。如果遗忘了您感兴趣的内容，请告诉我。

CATEGORIES_TO_TRAIN = ["apple", "dog", "cat"]

def createTFExample(img):
    imageFormat = ""
    if img.format == 'JPEG':
        imageFormat = b'jpeg'
    elif img.format == 'PNG':
        imageFormat = b'png'
    else:
        print 'Unknown Image format %s' %(img.format,)
        return None

    width, height = img.size
    filename = str(img.filename)
    encodedImageData = img.bytesIO

    xmins = []
    xmaxs = []
    ymins = []
    ymaxs = []

    for annotation in img.annotations:
        xmins.append((annotation.left / width))
        xmaxs.append((annotation.left + annotation.width) / width)
        ymins.append((annotation.top / height))
        ymaxs.append((annotation.top + annotation.height) / height)

    #we might have some images in our dataset, which don't have a annotation, skip those
    if((len(xmins) == 0) or (len(xmaxs) == 0) or (len(ymins) == 0) or (len(ymaxs) == 0)):
        return None

    label = [img.label.encode('utf8')]
    classes = [(CATEGORIES_TO_TRAIN.index(img.label) + 1)] #class indexes start with 1


    tf_example = tf.train.Example(features=tf.train.Features(feature={
      'image/height': dataset_util.int64_feature(height),
      'image/width': dataset_util.int64_feature(width),
      'image/filename': dataset_util.bytes_feature(filename),
      'image/source_id': dataset_util.bytes_feature(filename),
      'image/encoded': dataset_util.bytes_feature(encodedImageData),
      'image/format': dataset_util.bytes_feature(imageFormat),
      'image/object/bbox/xmin': dataset_util.float_list_feature(xmins),
      'image/object/bbox/xmax': dataset_util.float_list_feature(xmaxs),
      'image/object/bbox/ymin': dataset_util.float_list_feature(ymins),
      'image/object/bbox/ymax': dataset_util.float_list_feature(ymaxs),
      'image/object/class/text': dataset_util.bytes_list_feature(label),
      'image/object/class/label': dataset_util.int64_list_feature(classes),
    }))
    return tf_example


def createTfRecordFile(images):
    writer = tf.python_io.TFRecordWriter(TFRECORD_OUTPUT_PATH)
    for img in images:
        t = createTFExample(img)
        if t is not None:
            writer.write(t.SerializeToString())

    writer.close()

任何帮助我指向正确方向的帮助都非常感谢！

Answer 1

我有一个类似的问题，但让label列表和classes列表具有相同的长度，边界框元素为我修复了它。

具体来说，在createTFExample()中，label = [img.label.encode('utf8')]和classes = [(CATEGORIES_TO_TRAIN.index(img.label) + 1)]中的元素应对应于边界框注释列表的元素：

xmins = []
xmaxs = []
ymins = []
ymaxs = []

for annotation in img.annotations:
    xmins.append((annotation.left / width))
    xmaxs.append((annotation.left + annotation.width) / width)
    ymins.append((annotation.top / height))
    ymaxs.append((annotation.top + annotation.height) / height)

从你的代码结构我假设每个img对象有一个对象类型，但在这种情况下，写

label = [img.label.encode('utf8')] * len(xmins)  
classes = [(CATEGORIES_TO_TRAIN.index(img.label) + 1)] * len(xmins)

或使用任何能够为您提供图像中对象数量的元素，以便标签＆amp;类和边界框列表具有相同的长度。
如果img对象中有多种类型的对象，则创建一个对象名称和类别ID列表，其中内部元素的索引与注释列表的索引匹配。

结果列表应如下所示：

xmins = [a_xmin, b_xmin, c_xmin]
ymins = [a_ymin, b_ymin, c_ymin]
xmaxs = [a_xmax, b_xmax, c_xmax]
ymaxs = [a_ymax, b_ymax, c_ymax]
labels = [a_label, b_label, c_label]
classes = [a_classid, b_classid, c_classid]

这让我的问题消失了，希望这有用！

Tensorflow对象检测因InvalidArgumentError而中止：indices [0] = 2不在[0,1）中

1 个答案: