Ejemplo n.º 1
0
def mean_squared_error_regressor(tensor_in, labels, weights, biases, name=None):
  """Returns prediction and loss for mean squared error regression."""
  with ops.op_scope([tensor_in, labels], name, "mean_squared_error_regressor"):
    predictions = nn.xw_plus_b(tensor_in, weights, biases)
    if len(labels.get_shape()) == 1 and len(predictions.get_shape()) == 2:
      predictions = array_ops_.squeeze(predictions, squeeze_dims=[1])
    return predictions, loss_ops.sum_of_squares(predictions, labels)
Ejemplo n.º 2
0
def mean_squared_error_regressor(tensor_in,
                                 labels,
                                 weights,
                                 biases,
                                 name=None):
    """Returns prediction and loss for mean squared error regression."""
    with ops.op_scope([tensor_in, labels], name,
                      "mean_squared_error_regressor"):
        predictions = nn.xw_plus_b(tensor_in, weights, biases)
        if len(labels.get_shape()) == 1 and len(predictions.get_shape()) == 2:
            predictions = array_ops_.squeeze(predictions, squeeze_dims=[1])
        return predictions, loss_ops.sum_of_squares(predictions, labels)