torch_ecg.utils.select_k

torch_ecg.utils.select_k(arr: ndarray, k: int | List[int] | ndarray, dim: int = -1, largest: bool = True, sorted: bool = True) → Tuple[ndarray, ndarray]

Select elements from an array along a specified axis of specific rankings.

Parameters:

• arr (array_like) – Input array.

• k (int or array_like) – Number of elements to retrieve. If k is a 1D array, it represents the specific rankings to retrieve. NOTE that the rankings are 0-based, for example the rankings of the top 3 elements are [0, 1, 2].

• dim (int, default -1) – Axis along which to operate. Default is -1 (the last axis).

• largest (bool, default True) – If True, find the largest elements, else find the smallest elements.

• sorted (bool, default True) – If True, the result is sorted. If False, the result is not sorted.

Returns:

• values (numpy.ndarray) – The selected values along each axis.

• indices (numpy.ndarray) – The indices of the selected values along each axis.

• .. note:: – For integer k, this function has the same functionality as torch.topk().