v1.2.0

Release 1.2.0

Major Features and Improvements

• ModelCard updates
  • Fields
    • Add model_details.path.
      • This is populated with the new ModelSource object, which stores a reference to a TFX PushedModel.
    • Add model_parameters.input_format_map and model_parameters.output_format_map.
      • These are key-value pairs, and are used to render inputs and outputs in tabular form. They can be used as an alternative to the singular model_parameters.input_format and model_parameters.output_format fields.
  • Functions
    • Add from_json().
• model_card_toolkit.source
  • This is a new submodule, and is responsible for ModelCardToolkit's' inputs (see TFX standard artifacts). It provides the following classes:
    • MlmdSource: Args to extract data from TFX artifacts in MLMD. Contains the mlmd_store and model_uri args. Previously, these were args to ModelCardToolkit.
    • Source: Args to extract data from TFX artifacts outside MLMD (by passing in a path to the artifact, or by passing in the artifact directly). Contains tfma, tfdv, and model args.

Bug fixes and other changes

• ModelCardToolkit
  • model_card.quantitative_analysis.performance_metrics is now populated when a tfma.EvalResult is found in MLMD store.
  • export_format() and update_model_card() now accept model_card_pb2.ModelCard's, in addition to model_card.ModelCard's.
• tfx_util
  • Add annotate_eval_result_metrics(), which appends PerformanceMetrics to a ModelCard based on a tfma.EvalResult.
  • Add read_stats_protos(), which returns dataset stats protos for all splits in the provided directory.
  • Add filter_metrics() to facilitate filtering out unwanted TFMA metrics in model cards.
  • Add filter_features() and tfx._util.read_stats_protos_and_filter_features() to facilitate filtering out unwanted TFDV features in model cards.
• PerformanceMetrics
  • Add confidence_interval field.

Breaking changes and Deprecations

• Replace ModelCardToolkit(output_dir, mlmd_store, model_uri) with ModelCardToolkit(output_dir, mlmd_source, source). See "Major Features and Improvements" above for details.
• Complete deprecation of ModelCardToolkit.update_model_card_json(). Users should migrate to ModelCardToolkit.update_model_card(), which uses a proto representation. Alternatively, users can use ModelCard.to_json() and ModelCard.from_json() to interact with JSON representations.