lightning_template.utils.optim
==============================

.. py:module:: lightning_template.utils.optim


Submodules
----------

.. toctree::
   :maxdepth: 1

   /autoapi/lightning_template/utils/optim/configure_optimizers/index
   /autoapi/lightning_template/utils/optim/keep_and_linearly_decay_lr_scheduler/index
   /autoapi/lightning_template/utils/optim/warmup_lr_scheduler/index


Classes
-------

.. autoapisummary::

   lightning_template.utils.optim.KeepAndLinearlyDecayLrScheduler
   lightning_template.utils.optim.WarmupScheduler


Functions
---------

.. autoapisummary::

   lightning_template.utils.optim.get_configure_optimizers_method


Package Contents
----------------

.. py:function:: get_configure_optimizers_method(optim_config)

.. py:class:: KeepAndLinearlyDecayLrScheduler(optimizer, keep_epochs, decay_epochs, last_epoch=-1, verbose=False)

   Bases: :py:obj:`torch.optim.lr_scheduler._LRScheduler`


   Keep lr first then decay learning rate linearly in optimizer.

   :param optimizer: Wrapped optimizer.
   :type optimizer: Optimizer
   :param keep_epochs: target keep epoch.
   :param decay_epochs: target decay epoch.


   .. py:attribute:: keep_epochs


   .. py:attribute:: decay_epochs


   .. py:method:: get_lr()

      Compute learning rate using chainable form of the scheduler.



.. py:class:: WarmupScheduler(*args, warmup_iters, warmup_ratio=0.1, warmup_mode='linear', **kwargs)

   Bases: :py:obj:`torch.optim.lr_scheduler._LRScheduler`


   Warm-up(increasing) learning rate in optimizer.

   :param optimizer: Wrapped optimizer.
   :type optimizer: Optimizer
   :param warmup_iters: target warm up epoch.


   .. py:attribute:: warmup_iters


   .. py:attribute:: warmup_ratio
      :value: 0.1



   .. py:attribute:: warmup_mode
      :value: 'linear'



   .. py:method:: get_lr()

      Compute learning rate using chainable form of the scheduler.