Python object_detection.protos.losses_pb2.ClassificationLoss() Examples

The following are 30 code examples of object_detection.protos.losses_pb2.ClassificationLoss(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module object_detection.protos.losses_pb2 , or try the search function .
Example #1
Source File: losses_builder.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def build_faster_rcnn_classification_loss(loss_config):
  """Builds a classification loss for Faster RCNN based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    return losses.WeightedSigmoidClassificationLoss()
  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        logit_scale=config.logit_scale)
  if loss_type == 'weighted_logits_softmax':
    config = loss_config.weighted_logits_softmax
    return losses.WeightedSoftmaxClassificationAgainstLogitsLoss(
        logit_scale=config.logit_scale)
  if loss_type == 'weighted_sigmoid_focal':
    config = loss_config.weighted_sigmoid_focal
    alpha = None
    if config.HasField('alpha'):
      alpha = config.alpha
    return losses.SigmoidFocalClassificationLoss(
        gamma=config.gamma,
        alpha=alpha)

  # By default, Faster RCNN second stage classifier uses Softmax loss
  # with anchor-wise outputs.
  config = loss_config.weighted_softmax
  return losses.WeightedSoftmaxClassificationLoss(
      logit_scale=config.logit_scale)
Example #2
Source File: losses_builder.py    From Elphas with Apache License 2.0 5 votes vote down vote up 
def _build_classification_loss(loss_config):
  """Builds a classification loss based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    return losses.WeightedSigmoidClassificationLoss()

  if loss_type == 'weighted_sigmoid_focal':
    config = loss_config.weighted_sigmoid_focal
    alpha = None
    if config.HasField('alpha'):
      alpha = config.alpha
    return losses.SigmoidFocalClassificationLoss(
        gamma=config.gamma,
        alpha=alpha)

  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        logit_scale=config.logit_scale)

  if loss_type == 'bootstrapped_sigmoid':
    config = loss_config.bootstrapped_sigmoid
    return losses.BootstrappedSigmoidClassificationLoss(
        alpha=config.alpha,
        bootstrap_type=('hard' if config.hard_bootstrap else 'soft'))

  raise ValueError('Empty loss config.')
Example #3
Source File: losses_builder.py    From MBMD with MIT License 5 votes vote down vote up 
def _build_classification_loss(loss_config):
  """Builds a classification loss based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    config = loss_config.weighted_sigmoid
    return losses.WeightedSigmoidClassificationLoss(
        anchorwise_output=config.anchorwise_output)

  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        anchorwise_output=config.anchorwise_output)

  if loss_type == 'bootstrapped_sigmoid':
    config = loss_config.bootstrapped_sigmoid
    return losses.BootstrappedSigmoidClassificationLoss(
        alpha=config.alpha,
        bootstrap_type=('hard' if config.hard_bootstrap else 'soft'),
        anchorwise_output=config.anchorwise_output)

  raise ValueError('Empty loss config.')
Example #4
Source File: losses_builder_test.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up 
def test_build_softmax_loss(self):
    losses_text_proto = """
      weighted_softmax {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #5
Source File: losses_builder_test.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up 
def test_build_softmax_loss_by_default(self):
    losses_text_proto = """
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #6
Source File: losses_builder_test.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up 
def test_build_sigmoid_loss(self):
    losses_text_proto = """
      weighted_sigmoid {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSigmoidClassificationLoss))
Example #7
Source File: losses_builder_test.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up 
def test_build_softmax_loss(self):
    losses_text_proto = """
      weighted_softmax {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #8
Source File: losses_builder_test.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up 
def test_build_softmax_loss_by_default(self):
    losses_text_proto = """
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #9
Source File: losses_builder.py    From AniSeg with Apache License 2.0 5 votes vote down vote up 
def build_faster_rcnn_classification_loss(loss_config):
  """Builds a classification loss for Faster RCNN based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    return losses.WeightedSigmoidClassificationLoss()
  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        logit_scale=config.logit_scale)

  # By default, Faster RCNN second stage classifier uses Softmax loss
  # with anchor-wise outputs.
  config = loss_config.weighted_softmax
  return losses.WeightedSoftmaxClassificationLoss(
      logit_scale=config.logit_scale)
Example #10
Source File: losses_builder.py    From AniSeg with Apache License 2.0 5 votes vote down vote up 
def _build_classification_loss(loss_config):
  """Builds a classification loss based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    return losses.WeightedSigmoidClassificationLoss()

  if loss_type == 'weighted_sigmoid_focal':
    config = loss_config.weighted_sigmoid_focal
    alpha = None
    if config.HasField('alpha'):
      alpha = config.alpha
    return losses.SigmoidFocalClassificationLoss(
        gamma=config.gamma,
        alpha=alpha)

  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        logit_scale=config.logit_scale)

  if loss_type == 'bootstrapped_sigmoid':
    config = loss_config.bootstrapped_sigmoid
    return losses.BootstrappedSigmoidClassificationLoss(
        alpha=config.alpha,
        bootstrap_type=('hard' if config.hard_bootstrap else 'soft'))

  raise ValueError('Empty loss config.')
Example #11
Source File: losses_builder_test.py    From AniSeg with Apache License 2.0 5 votes vote down vote up 
def test_build_sigmoid_loss(self):
    losses_text_proto = """
      weighted_sigmoid {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSigmoidClassificationLoss))
Example #12
Source File: losses_builder_test.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up 
def test_build_sigmoid_loss(self):
    losses_text_proto = """
      weighted_sigmoid {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSigmoidClassificationLoss))
Example #13
Source File: losses_builder_test.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def test_build_sigmoid_focal_loss(self):
    losses_text_proto = """
      weighted_sigmoid_focal {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(
        isinstance(classification_loss,
                   losses.SigmoidFocalClassificationLoss))
Example #14
Source File: losses_builder_test.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def test_build_softmax_loss(self):
    losses_text_proto = """
      weighted_softmax {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #15
Source File: losses_builder_test.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def test_build_sigmoid_loss(self):
    losses_text_proto = """
      weighted_sigmoid {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSigmoidClassificationLoss))
Example #16
Source File: losses_builder_test.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def test_build_softmax_loss_by_default(self):
    losses_text_proto = """
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #17
Source File: losses_builder_test.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def test_build_softmax_loss_by_default(self):
    losses_text_proto = """
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #18
Source File: losses_builder_test.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def test_build_logits_softmax_loss(self):
    losses_text_proto = """
      weighted_logits_softmax {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(
        isinstance(classification_loss,
                   losses.WeightedSoftmaxClassificationAgainstLogitsLoss))
Example #19
Source File: losses_builder_test.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def test_build_softmax_loss(self):
    losses_text_proto = """
      weighted_softmax {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #20
Source File: losses_builder_test.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def test_build_sigmoid_loss(self):
    losses_text_proto = """
      weighted_sigmoid {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSigmoidClassificationLoss))
Example #21
Source File: losses_builder.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def build_faster_rcnn_classification_loss(loss_config):
  """Builds a classification loss for Faster RCNN based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    return losses.WeightedSigmoidClassificationLoss()
  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        logit_scale=config.logit_scale)
  if loss_type == 'weighted_logits_softmax':
    config = loss_config.weighted_logits_softmax
    return losses.WeightedSoftmaxClassificationAgainstLogitsLoss(
        logit_scale=config.logit_scale)
  if loss_type == 'weighted_sigmoid_focal':
    config = loss_config.weighted_sigmoid_focal
    alpha = None
    if config.HasField('alpha'):
      alpha = config.alpha
    return losses.SigmoidFocalClassificationLoss(
        gamma=config.gamma,
        alpha=alpha)

  # By default, Faster RCNN second stage classifier uses Softmax loss
  # with anchor-wise outputs.
  config = loss_config.weighted_softmax
  return losses.WeightedSoftmaxClassificationLoss(
      logit_scale=config.logit_scale)
Example #22
Source File: losses_builder.py    From hands-detection with MIT License 5 votes vote down vote up 
def _build_classification_loss(loss_config):
  """Builds a classification loss based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    config = loss_config.weighted_sigmoid
    return losses.WeightedSigmoidClassificationLoss(
        anchorwise_output=config.anchorwise_output)

  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        anchorwise_output=config.anchorwise_output)

  if loss_type == 'bootstrapped_sigmoid':
    config = loss_config.bootstrapped_sigmoid
    return losses.BootstrappedSigmoidClassificationLoss(
        alpha=config.alpha,
        bootstrap_type=('hard' if config.hard_bootstrap else 'soft'),
        anchorwise_output=config.anchorwise_output)

  raise ValueError('Empty loss config.')
Example #23
Source File: losses_builder.py    From moveo_ros with MIT License 5 votes vote down vote up 
def _build_classification_loss(loss_config):
  """Builds a classification loss based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    config = loss_config.weighted_sigmoid
    return losses.WeightedSigmoidClassificationLoss(
        anchorwise_output=config.anchorwise_output)

  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        anchorwise_output=config.anchorwise_output)

  if loss_type == 'bootstrapped_sigmoid':
    config = loss_config.bootstrapped_sigmoid
    return losses.BootstrappedSigmoidClassificationLoss(
        alpha=config.alpha,
        bootstrap_type=('hard' if config.hard_bootstrap else 'soft'),
        anchorwise_output=config.anchorwise_output)

  raise ValueError('Empty loss config.')
Example #24
Source File: losses_builder_test.py    From BMW-TensorFlow-Training-GUI with Apache License 2.0 5 votes vote down vote up 
def test_build_logits_softmax_loss(self):
    losses_text_proto = """
      weighted_logits_softmax {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(
        isinstance(classification_loss,
                   losses.WeightedSoftmaxClassificationAgainstLogitsLoss))
Example #25
Source File: losses_builder_test.py    From BMW-TensorFlow-Training-GUI with Apache License 2.0 5 votes vote down vote up 
def test_build_softmax_loss(self):
    losses_text_proto = """
      weighted_softmax {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSoftmaxClassificationLoss))
Example #26
Source File: losses_builder_test.py    From BMW-TensorFlow-Training-GUI with Apache License 2.0 5 votes vote down vote up 
def test_build_sigmoid_loss(self):
    losses_text_proto = """
      weighted_sigmoid {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSigmoidClassificationLoss))
Example #27
Source File: losses_builder.py    From BMW-TensorFlow-Training-GUI with Apache License 2.0 5 votes vote down vote up 
def build_faster_rcnn_classification_loss(loss_config):
  """Builds a classification loss for Faster RCNN based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    return losses.WeightedSigmoidClassificationLoss()
  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        logit_scale=config.logit_scale)
  if loss_type == 'weighted_logits_softmax':
    config = loss_config.weighted_logits_softmax
    return losses.WeightedSoftmaxClassificationAgainstLogitsLoss(
        logit_scale=config.logit_scale)

  # By default, Faster RCNN second stage classifier uses Softmax loss
  # with anchor-wise outputs.
  config = loss_config.weighted_softmax
  return losses.WeightedSoftmaxClassificationLoss(
      logit_scale=config.logit_scale)
Example #28
Source File: losses_builder_test.py    From ros_tensorflow with Apache License 2.0 5 votes vote down vote up 
def test_build_logits_softmax_loss(self):
    losses_text_proto = """
      weighted_logits_softmax {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(
        isinstance(classification_loss,
                   losses.WeightedSoftmaxClassificationAgainstLogitsLoss))
Example #29
Source File: losses_builder.py    From DOTA_models with Apache License 2.0 5 votes vote down vote up 
def _build_classification_loss(loss_config):
  """Builds a classification loss based on the loss config.

  Args:
    loss_config: A losses_pb2.ClassificationLoss object.

  Returns:
    Loss based on the config.

  Raises:
    ValueError: On invalid loss_config.
  """
  if not isinstance(loss_config, losses_pb2.ClassificationLoss):
    raise ValueError('loss_config not of type losses_pb2.ClassificationLoss.')

  loss_type = loss_config.WhichOneof('classification_loss')

  if loss_type == 'weighted_sigmoid':
    config = loss_config.weighted_sigmoid
    return losses.WeightedSigmoidClassificationLoss(
        anchorwise_output=config.anchorwise_output)

  if loss_type == 'weighted_softmax':
    config = loss_config.weighted_softmax
    return losses.WeightedSoftmaxClassificationLoss(
        anchorwise_output=config.anchorwise_output)

  if loss_type == 'bootstrapped_sigmoid':
    config = loss_config.bootstrapped_sigmoid
    return losses.BootstrappedSigmoidClassificationLoss(
        alpha=config.alpha,
        bootstrap_type=('hard' if config.hard_bootstrap else 'soft'),
        anchorwise_output=config.anchorwise_output)

  raise ValueError('Empty loss config.')
Example #30
Source File: losses_builder_test.py    From ros_tensorflow with Apache License 2.0 5 votes vote down vote up 
def test_build_sigmoid_loss(self):
    losses_text_proto = """
      weighted_sigmoid {
      }
    """
    losses_proto = losses_pb2.ClassificationLoss()
    text_format.Merge(losses_text_proto, losses_proto)
    classification_loss = losses_builder.build_faster_rcnn_classification_loss(
        losses_proto)
    self.assertTrue(isinstance(classification_loss,
                               losses.WeightedSigmoidClassificationLoss))