Skip to content

Utilities for passing TensorFlow-related metadata between tools

License

Notifications You must be signed in to change notification settings

tensorflow/metadata

Repository files navigation

TensorFlow Metadata

Python PyPI

TensorFlow Metadata provides standard representations for metadata that are useful when training machine learning models with TensorFlow.

The metadata serialization formats include:

  • A schema describing tabular data (e.g., tf.Examples).
  • A collection of summary statistics over such datasets.
  • A problem statement quantifying the objectives of a model.

The metadata may be produced by hand or automatically during input data analysis, and may be consumed for data validation, exploration, and transformation.