model_card_toolkit.TfmaSource

Stay organized with collections Save and categorize content based on your preferences.

Source to extract TFMA eval result data for a model card.

model_card_toolkit.TfmaSource(
    eval_result_paths: List[str] = dataclasses.field(default_factory=list),
    file_format: Optional[str] = '',
    model_evaluation_artifacts: List[metadata_store_pb2.Artifact] = dataclasses.field(default_factory=list),
    metrics_include: List[str] = dataclasses.field(default_factory=list),
    metrics_exclude: List[str] = dataclasses.field(default_factory=list)
)

Provide exactly one of eval_result_paths or model_evaluation_artifacts.

eval_result_paths The paths to the eval result output from TensorFlow Model Analysis or TFX Evaluator.
file_format Optional file extension to filter eval result files by.
model_evaluation_artifacts The MLMD artifacts from TensorFlow Model Analysis or TFX Evaluator.
metrics_include The list of metric names to include in the model card. By default, all metrics are included. Mutually exclusive with metrics_exclude.
metrics_exclude The list of metric names to exclude in the model card. By default, no metrics are excluded. Mutually exclusive with metrics_include.

Methods

__eq__

__eq__(
    other
)

file_format ''