tf.estimator.export.EvalOutput

Represents the output of a supervised eval process.

Inherits From: ExportOutput

This class generates the appropriate signature def for exporting eval output by type-checking and wrapping loss, predictions, and metrics values.

loss dict of Tensors or single Tensor representing calculated loss.
predictions dict of Tensors or single Tensor representing model predictions.
metrics Dict of metric results keyed by name. The values of the dict can be one of the following: (1) instance of Metric class. (2) (metric_value, update_op) tuples, or a single tuple. metric_value must be a Tensor, and update_op must be a Tensor or Op.

ValueError if any of the outputs' dict keys are not strings or tuples of strings or the values are not Tensors (or Operations in the case of update_op).

loss

metrics

predictions

Methods

as_signature_def

View source

Generate a SignatureDef proto for inclusion in a MetaGraphDef.

The SignatureDef will specify outputs as described in this ExportOutput, and will use the provided receiver_tensors as inputs.

Args
receiver_tensors a Tensor, or a dict of string to Tensor, specifying input nodes that will be fed.

LOSS_NAME 'loss'
METRICS_NAME 'metrics'
METRIC_UPDATE_SUFFIX 'update_op'
METRIC_VALUE_SUFFIX 'value'
PREDICTIONS_NAME 'predictions'