Change docstring type callable to Callable for consistency (#82487)

### Description

Across PyTorch's docstrings, both `callable` and `Callable` for variable types. The Callable should be capitalized as we are referring to the `Callable` type, and not the Python `callable()` function.

### Testing

There shouldn't be any testing required.

Pull Request resolved: https://github.com/pytorch/pytorch/pull/82487
Approved by: https://github.com/albanD

torch/contrib/_tensorboard_vis.py

@@ -57,7 +57,7 @@ def visualize_graph_executor(state, name_prefix, pb_graph, inline_graph):
         state (GraphExecutor or GraphExecutorState): GraphExecutor to display.
         name_prefix (str): Name prefix of the containing subgraph.
         pb_graph (GraphDef): graph to append to.
-        inline_graph (callable): a function that handles setting up a value_map,
+        inline_graph (Callable): a function that handles setting up a value_map,
             so that some graphs in here can be inlined. This is necessary, because
             this will simply be `visualize` for the top-level GraphExecutor,
             or `inline_graph` for all nested ones.

torch/distributed/_shard/sharded_optim/api.py

@@ -61,7 +61,7 @@ class ShardedOptimizer(optim.Optimizer):
         r"""Performs a single optimization step (parameter update).
 
         Args:
-            closure (callable): A closure that reevaluates the model and
+            closure (Callable): A closure that reevaluates the model and
                 returns the loss. Optional for most optimizers.
 
         .. note::

torch/distributed/algorithms/_quantization/quantization.py

@@ -95,11 +95,11 @@ def auto_quantize(func, qtype, quant_loss=None):
         . all_gather, all_to_all collective ops
     Note: BFP16 only supports 2D tensors.
     Args:
-        func (callable): A function representing collective operations.
+        func (Callable): A function representing collective operations.
         qtype (QuantType): Quantization method
         quant_loss (float, optional): This can be used to improve accuracy in the dequantization.
     Returns:
-        (callable): the same collective as func but enables automatic quantization/dequantization.
+        (Callable): the same collective as func but enables automatic quantization/dequantization.
     """
     @functools.wraps(func)
     def wrapper(*args, **kwargs):

torch/distributed/distributed_c10d.py

@@ -1073,7 +1073,7 @@ class P2POp(object):
     ``batch_isend_irecv`` for point-to-point communications.
 
     Args:
-        op (callable): A function to send data to or receive data from a peer process.
+        op (Callable): A function to send data to or receive data from a peer process.
             The type of ``op`` is either ``torch.distributed.isend`` or
             ``torch.distributed.irecv``.
         tensor (Tensor): Tensor to send or receive.

torch/distributed/fsdp/fully_sharded_data_parallel.py

@@ -4128,7 +4128,7 @@ class FullyShardedDataParallel(nn.Module):
                             peers to communicate with next in `GossipGrad <https://arxiv.org/abs/1803.05880>`_, etc.
                             It is locally stored by each worker
                             and shared by all the gradient tensors on the worker.
-            hook (callable): Callable with the following signature:
+            hook (Callable): Callable with the following signature:
                             ``hook: Callable[torch.Tensor] -> None``:
                             This function takes in a Python tensor, which represents
                             the full, flattened, unsharded gradient with respect to all variables

torch/distributed/optim/zero_redundancy_optimizer.py

@@ -984,7 +984,7 @@ class ZeroRedundancyOptimizer(Optimizer, Joinable):
                 If the argument itself is ``None``, then all parameters are
                 updated, and the gradients are assumed to be already populated.
                 (default: ``None``)
-            closure (callable): a closure that re-evaluates the model and
+            closure (Callable): a closure that re-evaluates the model and
                 returns the loss; optional for most optimizers and should be
                 ``None`` if ``gradients`` is not ``None``; (default: ``None``)
         Returns:
@@ -1043,7 +1043,7 @@ class ZeroRedundancyOptimizer(Optimizer, Joinable):
         Performs a single optimizer step and syncs parameters across all ranks.
 
         Arguments:
-            closure (callable): a closure that re-evaluates the model and
+            closure (Callable): a closure that re-evaluates the model and
                 returns the loss; optional for most optimizers.
         Returns:
             Optional loss depending on the underlying local optimizer.

torch/distributed/rpc/api.py

@@ -528,7 +528,7 @@ def remote(to, func, args=None, kwargs=None, timeout=UNSET_RPC_TIMEOUT):
 
     Args:
         to (str or WorkerInfo or int): name/rank/``WorkerInfo`` of the destination worker.
-        func (callable): a callable function, such as Python callables, builtin
+        func (Callable): a callable function, such as Python callables, builtin
                          operators (e.g. :meth:`~torch.add`) and annotated
                          TorchScript functions.
         args (tuple): the argument tuple for the ``func`` invocation.
@@ -736,7 +736,7 @@ def rpc_sync(to, func, args=None, kwargs=None, timeout=UNSET_RPC_TIMEOUT):
 
     Args:
         to (str or WorkerInfo or int): name/rank/``WorkerInfo`` of the destination worker.
-        func (callable): a callable function, such as Python callables, builtin
+        func (Callable): a callable function, such as Python callables, builtin
                          operators (e.g. :meth:`~torch.add`) and annotated
                          TorchScript functions.
         args (tuple): the argument tuple for the ``func`` invocation.
@@ -810,7 +810,7 @@ def rpc_async(to, func, args=None, kwargs=None, timeout=UNSET_RPC_TIMEOUT):
 
     Args:
         to (str or WorkerInfo or int): name/rank/``WorkerInfo`` of the destination worker.
-        func (callable): a callable function, such as Python callables, builtin
+        func (Callable): a callable function, such as Python callables, builtin
                          operators (e.g. :meth:`~torch.add`) and annotated
                          TorchScript functions.
         args (tuple): the argument tuple for the ``func`` invocation.

torch/distributions/constraint_registry.py

@@ -98,7 +98,7 @@ class ConstraintRegistry(object):
             constraint (subclass of :class:`~torch.distributions.constraints.Constraint`):
                 A subclass of :class:`~torch.distributions.constraints.Constraint`, or
                 a singleton object of the desired class.
-            factory (callable): A callable that inputs a constraint object and returns
+            factory (Callable): A callable that inputs a constraint object and returns
                 a  :class:`~torch.distributions.transforms.Transform` object.
         """
         # Support use as decorator.

torch/distributions/constraints.py

@@ -159,7 +159,7 @@ class _DependentProperty(property, _Dependent):
                 return constraints.interval(self.low, self.high)
 
     Args:
-        fn (callable): The function to be decorated.
+        fn (Callable): The function to be decorated.
         is_discrete (bool): Optional value of ``.is_discrete`` in case this
             can be computed statically. If not provided, access to the
             ``.is_discrete`` attribute will raise a NotImplementedError.

torch/jit/_script.py

@@ -1053,7 +1053,7 @@ def script(obj, optimize=None, _frames_up=0, _rcb=None,
      and as a decorator ``@torch.jit.script`` for :ref:`torchscript-classes` and functions.
 
     Args:
-        obj (callable, class, or nn.Module):  The ``nn.Module``, function, class type,
+        obj (Callable, class, or nn.Module):  The ``nn.Module``, function, class type,
                                                   dictionary, or list to compile.
         example_inputs (Union[List[Tuple], Dict[Callable, List[Tuple]], None]): Provide example inputs
             to annotate the arguments for a function or ``nn.Module``.

torch/nn/modules/loss.py

@@ -1528,7 +1528,7 @@ class TripletMarginWithDistanceLoss(_Loss):
     loss for input tensors using the :math:`l_p` distance as the distance function.
 
     Args:
-        distance_function (callable, optional): A nonnegative, real-valued function that
+        distance_function (Callable, optional): A nonnegative, real-valued function that
             quantifies the closeness of two tensors. If not specified,
             `nn.PairwiseDistance` will be used.  Default: ``None``
         margin (float, optional): A nonnegative margin representing the minimum difference

torch/nn/parallel/distributed.py

@@ -1324,7 +1324,7 @@ class DistributedDataParallel(Module, Joinable):
                                 _BufferCommHookLocation.POST_FORWARD means that the
                                 hook will run _after_ the forward pass.
 
-                hook (callable): Callable with the following signature:
+                hook (Callable): Callable with the following signature:
                              ``hook(state: object, bucket: dist.GradBucket) -> torch.futures.Future[torch.Tensor]``:
 
                 NOTE: To maximize performance, users can return a
@@ -1364,7 +1364,7 @@ class DistributedDataParallel(Module, Joinable):
 
                             It is locally stored by each worker
                             and shared by all the gradient tensors on the worker.
-            hook (callable): Callable with the following signature:
+            hook (Callable): Callable with the following signature:
                              ``hook(state: object, bucket: dist.GradBucket) -> torch.futures.Future[torch.Tensor]``:
 
                              This function is called once the bucket is ready. The

torch/optim/adadelta.py

@@ -79,7 +79,7 @@ class Adadelta(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/adagrad.py

@@ -125,7 +125,7 @@ class Adagrad(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/adam.py

@@ -108,7 +108,7 @@ class Adam(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         self._cuda_graph_capture_health_check()

torch/optim/adamax.py

@@ -83,7 +83,7 @@ class Adamax(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/adamw.py

@@ -109,7 +109,7 @@ class AdamW(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         self._cuda_graph_capture_health_check()

torch/optim/asgd.py

@@ -65,7 +65,7 @@ class ASGD(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/lbfgs.py

@@ -285,7 +285,7 @@ class LBFGS(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable): A closure that reevaluates the model
+            closure (Callable): A closure that reevaluates the model
                 and returns the loss.
         """
         assert len(self.param_groups) == 1

torch/optim/nadam.py

@@ -93,7 +93,7 @@ class NAdam(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/optimizer.py

@@ -272,7 +272,7 @@ class Optimizer(object):
         r"""Performs a single optimization step (parameter update).
 
         Args:
-            closure (callable): A closure that reevaluates the model and
+            closure (Callable): A closure that reevaluates the model and
                 returns the loss. Optional for most optimizers.
 
         .. note::

torch/optim/radam.py

@@ -91,7 +91,7 @@ class RAdam(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/rmsprop.py

@@ -97,7 +97,7 @@ class RMSprop(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/rprop.py

@@ -74,7 +74,7 @@ class Rprop(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/sgd.py

@@ -119,7 +119,7 @@ class SGD(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/optim/sparse_adam.py

@@ -58,7 +58,7 @@ class SparseAdam(Optimizer):
         """Performs a single optimization step.
 
         Args:
-            closure (callable, optional): A closure that reevaluates the model
+            closure (Callable, optional): A closure that reevaluates the model
                 and returns the loss.
         """
         loss = None

torch/profiler/profiler.py

@@ -281,9 +281,9 @@ class profile(_KinetoProfile):
         activities (iterable): list of activity groups (CPU, CUDA) to use in profiling, supported values:
             ``torch.profiler.ProfilerActivity.CPU``, ``torch.profiler.ProfilerActivity.CUDA``.
             Default value: ProfilerActivity.CPU and (when available) ProfilerActivity.CUDA.
-        schedule (callable): callable that takes step (int) as a single parameter and returns
+        schedule (Callable): callable that takes step (int) as a single parameter and returns
             ``ProfilerAction`` value that specifies the profiler action to perform at each step.
-        on_trace_ready (callable): callable that is called at each step when ``schedule``
+        on_trace_ready (Callable): callable that is called at each step when ``schedule``
             returns ``ProfilerAction.RECORD_AND_SAVE`` during the profiling.
         record_shapes (bool): save information about operator's input shapes.
         profile_memory (bool): track tensor memory allocation/deallocation.

torch/testing/_internal/common_utils.py

@@ -294,7 +294,7 @@ class parametrize(_TestParametrizer):
         arg_str (str): String of arg names separate by commas (e.g. "x,y").
         arg_values (iterable): Iterable of arg values (e.g. range(10)) or
             tuples of arg values (e.g. [(1, 2), (3, 4)]).
-        name_fn (callable): Optional function that takes in parameters and returns subtest name.
+        name_fn (Callable): Optional function that takes in parameters and returns subtest name.
     """
     def __init__(self, arg_str, arg_values, name_fn=None):
         self.arg_names: List[str] = [s.strip() for s in arg_str.split(',')]

torch/utils/data/dataloader.py

@@ -146,7 +146,7 @@ class DataLoader(Generic[T_co]):
         num_workers (int, optional): how many subprocesses to use for data
             loading. ``0`` means that the data will be loaded in the main process.
             (default: ``0``)
-        collate_fn (callable, optional): merges a list of samples to form a
+        collate_fn (Callable, optional): merges a list of samples to form a
             mini-batch of Tensor(s).  Used when using batched loading from a
             map-style dataset.
         pin_memory (bool, optional): If ``True``, the data loader will copy Tensors
@@ -159,7 +159,7 @@ class DataLoader(Generic[T_co]):
             will be smaller. (default: ``False``)
         timeout (numeric, optional): if positive, the timeout value for collecting a batch
             from workers. Should always be non-negative. (default: ``0``)
-        worker_init_fn (callable, optional): If not ``None``, this will be called on each
+        worker_init_fn (Callable, optional): If not ``None``, this will be called on each
             worker subprocess with the worker id (an int in ``[0, num_workers - 1]``) as
             input, after seeding and before data loading. (default: ``None``)
         generator (torch.Generator, optional): If not ``None``, this RNG will be used