
    i                     x    d dl mZ d dlZd dlmZ d dlmZ 	 ddedededeeef   d	ed
efdZddeded	ed
efdZ	y)    )UnionN)Tensor)_r2_score_updatesum_squared_obssum_obssum_squared_errornum_obssquaredreturnc                     t        j                  |j                        j                  }|t        j                  | ||z  |z  z
  |      z  }|st        j
                  |      }t        j                  |      S )a  Computes Relative Squared Error.

    Args:
        sum_squared_obs: Sum of square of all observations
        sum_obs: Sum of all observations
        sum_squared_error: Residual sum of squares
        num_obs: Number of predictions or observations
        squared: Returns RRSE value if set to False.

    Example:
        >>> target = torch.tensor([[0.5, 1], [-1, 1], [7, -6]])
        >>> preds = torch.tensor([[0, 2], [-1, 2], [8, -5]])
        >>> # RSE uses the same update function as R2 score.
        >>> sum_squared_obs, sum_obs, rss, num_obs = _r2_score_update(preds, target)
        >>> _relative_squared_error_compute(sum_squared_obs, sum_obs, rss, num_obs, squared=True)
        tensor(0.0632)

    )min)torchfinfodtypeepsclampsqrtmean)r   r   r   r	   r
   epsilonrses          {/Volumes/fast/ai/experiments/voice-extract-mac/.venv/lib/python3.12/site-packages/torchmetrics/functional/regression/rse.py_relative_squared_error_computer      sc    2 kk+11266G
ekk/Gg<MPW<W*W]de
eCjjo::c?    predstargetc                 D    t        | |      \  }}}}t        |||||      S )a"  Computes the relative squared error (RSE).

    .. math:: \text{RSE} = \frac{\sum_i^N(y_i - \hat{y_i})^2}{\sum_i^N(y_i - \overline{y})^2}

    Where :math:`y` is a tensor of target values with mean :math:`\overline{y}`, and
    :math:`\hat{y}` is a tensor of predictions.

    If `preds` and `targets` are 2D tensors, the RSE is averaged over the second dim.

    Args:
        preds: estimated labels
        target: ground truth labels
        squared: returns RRSE value if set to False
    Return:
        Tensor with RSE

    Example:
        >>> from torchmetrics.functional.regression import relative_squared_error
        >>> target = torch.tensor([3, -0.5, 2, 7])
        >>> preds = torch.tensor([2.5, 0.0, 2, 8])
        >>> relative_squared_error(preds, target)
        tensor(0.0514)

    )r
   )r   r   )r   r   r
   r   r   rssr	   s          r   relative_squared_errorr   6   s-    2 .>eV-L*OWc7*?GS'[bccr   )T)
typingr   r   r   %torchmetrics.functional.regression.r2r   intboolr   r    r   r   <module>r$      s       B   3;	
  @d& d& d4 dSY dr   