Missed TensorFlow Dev Summit? Check out the video playlist. Watch recordings


View source on GitHub

Infers schema from the input statistics.



  • statistics: A DatasetFeatureStatisticsList protocol buffer. Schema inference is currently supported only for lists with a single DatasetFeatureStatistics proto or lists with multiple DatasetFeatureStatistics protos corresponding to data slices that include the default slice (i.e., the slice with all examples). If a list with multiple DatasetFeatureStatistics protos is used, this function will infer the schema from the statistics corresponding to the default slice.
  • infer_feature_shape: A boolean to indicate if shape of the features need to be inferred from the statistics.
  • max_string_domain_size: Maximum size of the domain of a string feature in order to be interpreted as a categorical feature.
  • schema_transformations: List of transformation functions to apply to the auto-inferred schema. Each transformation function should take the schema and statistics as input and should return the transformed schema. The transformations are applied in the order provided in the list.


A Schema protocol buffer.


  • TypeError: If the input argument is not of the expected type.
  • ValueError: If the input statistics proto contains multiple datasets, none of which corresponds to the default slice.