noether.data.stats
==================

.. py:module:: noether.data.stats


Classes
-------

.. autoapisummary::

   noether.data.stats.RunningMoments
   noether.data.stats.RunningStats


Module Contents
---------------

.. py:class:: RunningMoments(log_scale = False)

   Bases: :py:obj:`torch.nn.Module`


   Calculates running moments of data (mean, variance, min, max) for data normalization purposes.

   This class implements Welford's online algorithm for computing mean and variance
   in a single pass, making it memory-efficient for large datasets.

   Initialize a RunningMoments instance.

   :param log_scale: Whether to apply log scaling to the data before calculating moments.
                     This affects all operations unless overridden.


   .. py:attribute:: log_scale
      :value: False



   .. py:method:: push_tensor(x, dim = 1, log_scale = None)

      Add a tensor to the moment calculation. Calculations are carried out in float64 to avoid numerical
      imprecisions.

      :param x: Tensor to push. Shape should be compatible with dim parameter.
      :param dim: Which dimension contains the feature dimension. For example:
                  - In a pointcloud (N, 3), dim=1 calculates statistics per coordinate (x, y, z)
                  - In an RGB image (B, C, H, W), dim=1 calculates per-channel statistics
      :param log_scale: Whether to apply log scaling to this tensor. If None, uses the class default.

      :raises ValueError: If tensor has invalid dimensions or is incompatible with previous data



   .. py:method:: push_scalar(x, log_scale = None)

      Add a scalar to the moment calculation. Calculations are carried out in float64.

      :param x: Scalar value to add to the moment calculations.
      :param log_scale: Whether to apply log scaling to this scalar. If None, uses the class default.

      :raises TypeError: If x is not a Python float



   .. py:property:: mean
      :type: Union[float, torch.Tensor]



   .. py:property:: var
      :type: Union[float, torch.Tensor]



   .. py:property:: std
      :type: Union[float, torch.Tensor]



   .. py:property:: min
      :type: Union[float, torch.Tensor]



   .. py:property:: max
      :type: Union[float, torch.Tensor]



   .. py:property:: count
      :type: int



   .. py:method:: print()

      Print the calculated statistics in a readable format.



   .. py:method:: dump()


   .. py:method:: reset()

      Reset the statistics, clearing all accumulated data.



   .. py:method:: is_empty()


.. py:class:: RunningStats(name = None)

   Calculates statistics of data (min, max, mean and variance) for data normalization purposes.


   .. py:attribute:: name
      :value: None



   .. py:method:: push_tensor(x, dim = 1)

      Add a tensor to the statistics. Calculations are carried out in float64 to avoid numerical imprecision.



   .. py:property:: min
      :type: torch.Tensor



   .. py:property:: max
      :type: torch.Tensor



   .. py:property:: mean
      :type: torch.Tensor



   .. py:property:: var
      :type: torch.Tensor



   .. py:property:: std
      :type: torch.Tensor



   .. py:property:: logmean
      :type: torch.Tensor



   .. py:property:: logvar
      :type: torch.Tensor



   .. py:property:: logstd
      :type: torch.Tensor