tf.compat.v1.saved_model.regression_signature_def

Creates regression signature from given examples and predictions.

This function produces signatures intended for use with the TensorFlow Serving Regress API (tensorflow_serving/apis/prediction_service.proto), and so constrains the input and output types to those allowed by TensorFlow Serving.

examples A string Tensor, expected to accept serialized tf.Examples.
predictions A float Tensor.

A regression-flavored signature_def.

ValueError If examples is None.