TensorFlow 2.0 Beta is available Learn more

tfma.evaluators.AnalysisTableEvaluator

tfma.evaluators.AnalysisTableEvaluator(
    key=constants.ANALYSIS_KEY,
    run_after=extractor.LAST_EXTRACTOR_STAGE_NAME,
    include=None,
    exclude=None
)

Defined in evaluators/analysis_table_evaluator.py.

Creates an Evaluator for returning Extracts data for analysis.

If both include and exclude are None then tfma.INPUT_KEY extracts will be excluded by default.

Args:

  • key: Name to use for key in Evaluation output.
  • run_after: Extractor to run after (None means before any extractors).
  • include: Keys of extracts to include in output. Keys starting with '_' are automatically filtered out at write time.
  • exclude: Keys of extracts to exclude from output.

Returns:

Evaluator for collecting analysis data. The output is stored under the key 'analysis'.

Raises:

  • ValueError: If both include and exclude are used.