flame.pytorch.metrics.base_metric

Module Contents

Classes

Metric
MetricList

Functions

_to_number(x)
_default_compute_fn(x)

flame.pytorch.metrics.base_metric._to_number(x)

Return type:

numbers.Number

flame.pytorch.metrics.base_metric._default_compute_fn(x)

Return type:

Any

class flame.pytorch.metrics.base_metric.Metric(name, output_transform, compute_fn=_default_compute_fn)

Parameters:

• name (Union[str, Sequence[str]]) –

• output_transform (Callable) –

• compute_fn (Callable) –

__call__(output)

Parameters:

output (dict) –

Return type:

dict

class flame.pytorch.metrics.base_metric.MetricList(metrics)

Parameters:

metrics (List[Metric]) –

__call__(output)

Parameters:

output (dict) –

Return type:

dict