ISubGVQA.utils.misc
===================

.. py:module:: ISubGVQA.utils.misc

.. autoapi-nested-parse::

   Misc functions, including distributed helpers.

   Mostly copy-paste from torchvision references.



Classes
-------

.. autoapisummary::

   ISubGVQA.utils.misc.SmoothedValue
   ISubGVQA.utils.misc.MetricLogger
   ISubGVQA.utils.misc.NestedTensor


Functions
---------

.. autoapisummary::

   ISubGVQA.utils.misc.all_gather
   ISubGVQA.utils.misc.reduce_dict
   ISubGVQA.utils.misc.get_sha
   ISubGVQA.utils.misc.collate_fn
   ISubGVQA.utils.misc._max_by_axis
   ISubGVQA.utils.misc.nested_tensor_from_tensor_list
   ISubGVQA.utils.misc.setup_for_distributed
   ISubGVQA.utils.misc.is_dist_avail_and_initialized
   ISubGVQA.utils.misc.get_world_size
   ISubGVQA.utils.misc.get_rank
   ISubGVQA.utils.misc.is_main_process
   ISubGVQA.utils.misc.save_on_master
   ISubGVQA.utils.misc.init_distributed_mode
   ISubGVQA.utils.misc.accuracy
   ISubGVQA.utils.misc.interpolate


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

.. py:class:: SmoothedValue(window_size=20, fmt=None)

   Bases: :py:obj:`object`


   Track a series of values and provide access to smoothed values over a
   window or the global series average.


   .. py:attribute:: deque


   .. py:attribute:: total
      :value: 0.0



   .. py:attribute:: count
      :value: 0



   .. py:attribute:: fmt
      :value: None



   .. py:method:: update(value, n=1)


   .. py:method:: synchronize_between_processes()

      Warning: does not synchronize the deque!



   .. py:property:: median


   .. py:property:: avg


   .. py:property:: global_avg


   .. py:property:: max


   .. py:property:: value


   .. py:method:: __str__()


.. py:function:: all_gather(data)

   Run all_gather on arbitrary picklable data (not necessarily tensors)
   Args:
       data: any picklable object
   Returns:
       list[data]: list of data gathered from each rank


.. py:function:: reduce_dict(input_dict, average=True)

   Args:
       input_dict (dict): all the values will be reduced
       average (bool): whether to do average or sum
   Reduce the values in the dictionary from all processes so that all processes
   have the averaged results. Returns a dict with the same fields as
   input_dict, after reduction.


.. py:class:: MetricLogger(delimiter='\t')

   Bases: :py:obj:`object`


   .. py:attribute:: meters


   .. py:attribute:: delimiter
      :value: '\t'



   .. py:method:: update(**kwargs)


   .. py:method:: __getattr__(attr)


   .. py:method:: __str__()


   .. py:method:: synchronize_between_processes()


   .. py:method:: add_meter(name, meter)


   .. py:method:: log_every(iterable, print_freq, header=None)


.. py:function:: get_sha()

.. py:function:: collate_fn(batch)

.. py:function:: _max_by_axis(the_list: List[List[int]]) -> List[int]

.. py:function:: nested_tensor_from_tensor_list(tensor_list: List[torch.Tensor])

.. py:class:: NestedTensor(tensors, mask: Optional[torch.Tensor])

   Bases: :py:obj:`object`


   .. py:attribute:: tensors


   .. py:attribute:: mask


   .. py:method:: to(device: Device) -> NestedTensor


   .. py:method:: decompose()


   .. py:method:: __repr__()


.. py:function:: setup_for_distributed(is_master)

   This function disables printing when not in master process


.. py:function:: is_dist_avail_and_initialized()

.. py:function:: get_world_size()

.. py:function:: get_rank()

.. py:function:: is_main_process()

.. py:function:: save_on_master(*args, **kwargs)

.. py:function:: init_distributed_mode(args)

.. py:function:: accuracy(output, target, topk=(1, ))

   Computes the precision@k for the specified values of k


.. py:function:: interpolate(input: torch.Tensor, size: Optional[List[int]] = None, scale_factor: Optional[float] = None, mode: str = 'nearest', align_corners: Optional[bool] = None) -> torch.Tensor

   Equivalent to nn.functional.interpolate, but with support for empty batch sizes.
   This will eventually be supported natively by PyTorch, and this
   class can go away.