cleaning of code

pytorch · HaoZeke · Jun 8, 2021 · Jun 10, 2021 · Jun 10, 2021 · Jun 17, 2021
commit f30497f040b051b81cdfc1a7249335eae7f801af
diff --git a/test/test_numpy_interop.py b/test/test_numpy_interop.py
@@ -161,8 +161,8 @@ def test_to_numpy_force_argument(self, device) -> None:
             for requires_grad in [False, True]:
                 for sparse in [False, True]:
                     for conj in [False, True]:
-                        data = [[1+2j, -2+3j], [-1-2j, 3-2j]]
-                        x = torch.tensor(data, requires_grad = requires_grad, device = device)    	
+                        data = [[1 + 2j, -2 + 3j], [-1 - 2j, 3 - 2j]]
+                        x = torch.tensor(data, requires_grad=requires_grad, device=device)
                         y = x
                         if sparse:
                             if requires_grad:

diff --git a/tools/autograd/templates/python_variable_methods.cpp b/tools/autograd/templates/python_variable_methods.cpp
@@ -800,11 +800,11 @@ static PyObject * THPVariable_numpy(PyObject* self, PyObject* args, PyObject* kw
   auto& self_ = THPVariable_Unpack(self);
   ParsedArgs<1> parsed_args;
   auto r = parser.parse(self, args, kwargs, parsed_args);
-  
+
   if (r.has_torch_function()) {
     return handle_torch_function(r, self, args, kwargs, THPVariableClass, "torch.Tensor");
   }
-  
+
   jit::tracer::warn("Converting a tensor to a NumPy array", jit::tracer::WARN_PYTHON_DATAFLOW);
   return torch::utils::tensor_to_numpy(self_, r.toBool(0));
   END_HANDLE_TH_ERRORS

diff --git a/torch/_tensor_docs.py b/torch/_tensor_docs.py
@@ -2848,15 +2848,15 @@ def callable(a, b) -> number
 If :attr:`force` is ``False``, tensor :attr:`self` and the
 returned :class:`ndarray` share the same underlying storage. Changes to
 :attr:`self` tensor will be reflected in the :class:`ndarray` and vice versa.
-If tensors storage is not CPU, or if :attr:`requires_grad` is ``True`` and 
-:attr:`GradMode` is ``True``, or tensor is sprase, or :attr:`is_conj()` 
+If tensors storage is not CPU, or if :attr:`requires_grad` is ``True`` and
+:attr:`GradMode` is ``True``, or tensor is sprase, or :attr:`is_conj()`
 is ``True``, it throws exception.
 
 If :attr:`force` is ``True`` it tries to resolve issues, that would throw exception.
-It calls :func:`torch.Tensor.detach`, :func:`torch.Tensor.cpu` and if :attr:`is_conj()` 
+It calls :func:`torch.Tensor.detach`, :func:`torch.Tensor.cpu` and if :attr:`is_conj()`
 is ``True`` it calls :func:`torch.Tensor.resolve_conj`.
 Therefore tensor :attr:`self` and the returned :class:`ndarray` can have separate
-underlying storage and changes to :attr:`self` may not be reflected 
+underlying storage and changes to :attr:`self` may not be reflected
 in the :class:`ndarray` and vice versa.
 
 Args: