5.1.1.1.1.7. FedEval.aggregator.trim

5.1.1.1.1.7.1. Module Contents

5.1.1.1.1.7.1.1. Functions

_trim_ndarrays(→ Iterable[numpy.ndarray])

trim_params(...)

Return the params after trimming a certain ratio of the extreme parameter values.
FedEval.aggregator.trim._trim_ndarrays(arrs: Iterable[numpy.ndarray], num_trim: int = 0) → Iterable[numpy.ndarray]
FedEval.aggregator.trim.trim_params(params: Iterable[FedEval.aggregator.ModelWeight.ModelWeights], ratio: float = 0.05) → Iterable[FedEval.aggregator.ModelWeight.ModelWeights]

Return the params after trimming a certain ratio of the extreme parameter values.

Parameters:

  • params (Iterable[ModelWeights]) – The weights of models, ordered like [params1, params2, …].

  • ratio (float, optional) – The ratio of extreme parameter values to trim. Should be between 0 and 0.5. Defaults to 0.05.

Raises:

ValueError – If trim_ratio is in [0, 0.5).

Returns:

trimmed params

Return type:

Iterable[ModelWeights]