agedi.diffusion.sdes.noise_schedules
====================================

.. py:module:: agedi.diffusion.sdes.noise_schedules


Classes
-------

.. autoapisummary::

   agedi.diffusion.sdes.noise_schedules.NoiseSchedule
   agedi.diffusion.sdes.noise_schedules.Linear
   agedi.diffusion.sdes.noise_schedules.Exponential
   agedi.diffusion.sdes.noise_schedules.Cosine


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

.. py:class:: NoiseSchedule(min: float, max: float)

   Bases: :py:obj:`abc.ABC`


   Abstract base class for diffusion noise schedules.

   A noise schedule defines a function ``f(t)`` that controls the noise level
   during the forward diffusion process, where ``t ∈ [0, 1]``.


   .. py:attribute:: min


   .. py:attribute:: max


   .. py:method:: get_hparams() -> Dict

      Return hyperparameters sufficient to reconstruct this noise schedule.

      Returns a dictionary with a ``_target_`` key plus ``min`` and ``max``.
      Subclasses can call ``super().get_hparams()`` and add their own params.

      :returns: Hyperparameter dictionary.
      :rtype: dict



   .. py:method:: f(t: torch.Tensor) -> torch.Tensor
      :abstractmethod:


      Returns the noise schedule value at time t.



   .. py:method:: fprime(t: torch.Tensor) -> torch.Tensor
      :abstractmethod:


      Returns the derivative of the noise schedule at time t.



   .. py:method:: fint(t: torch.Tensor) -> torch.Tensor
      :abstractmethod:


      Return the integral of the noise schedule at time t



   .. py:method:: df2dt(t: torch.Tensor) -> torch.Tensor

      Return the time derivative of f(t)² at time *t*.

      Computed as ``2 * f(t) * f'(t)``.



.. py:class:: Linear(min: float, max: float)

   Bases: :py:obj:`NoiseSchedule`


   Linear noise schedule: ``f(t) = min + (max - min) * t``.


   .. py:method:: f(t: torch.Tensor) -> torch.Tensor

      Evaluate the noise schedule at time *t*.



   .. py:method:: fprime(t: torch.Tensor) -> torch.Tensor

      Return the derivative of the noise schedule at time *t*.



   .. py:method:: fint(t: torch.Tensor) -> torch.Tensor

      Return the integral of the noise schedule from 0 to *t*.



.. py:class:: Exponential(min: float, max: float)

   Bases: :py:obj:`NoiseSchedule`


   Exponential noise schedule: ``f(t) = min * (max/min)^t``.


   .. py:method:: f(t: torch.Tensor) -> torch.Tensor

      Evaluate the noise schedule at time *t*.



   .. py:method:: fprime(t: torch.Tensor) -> torch.Tensor

      Return the derivative of the noise schedule at time *t*.



   .. py:method:: fint(t: torch.Tensor) -> torch.Tensor

      Return the integral of the noise schedule from 0 to *t*.



.. py:class:: Cosine(min: float, max: float)

   Bases: :py:obj:`NoiseSchedule`


   Cosine noise schedule: ``f(t) = min + (max - min) * (1 - cos(πt)) / 2``.


   .. py:method:: f(t: torch.Tensor) -> torch.Tensor

      Evaluate the noise schedule at time *t*.



   .. py:method:: fprime(t: torch.Tensor) -> torch.Tensor

      Return the derivative of the noise schedule at time *t*.



   .. py:method:: fint(t: torch.Tensor) -> torch.Tensor

      Return the integral of the noise schedule from 0 to *t*.