spelling check pt2 on docstrings in opacus folder

opacus/accountants/accountant.py

@@ -72,7 +72,7 @@ def get_optimizer_hook_fn(
         """
         Returns a callback function which can be used to attach to DPOptimizer
         Args:
-            sample_rate: Expected samping rate used for accounting
+            sample_rate: Expected sampling rate used for accounting
         """
 
         def hook_fn(optim: DPOptimizer):
@@ -88,7 +88,7 @@ def hook_fn(optim: DPOptimizer):
 
     def state_dict(self, destination: T_state_dict = None) -> T_state_dict:
         """
-        Retruns a dictionary containing the state of the accountant.
+        Returns a dictionary containing the state of the accountant.
         Args:
             destination: a mappable object to populate the current state_dict into.
                 If this arg is None, an OrderedDict is created and populated.

opacus/optimizers/ddp_perlayeroptimizer.py

@@ -67,7 +67,7 @@ def __init__(
 class DistributedPerLayerOptimizer(DPOptimizer):
     """
     :class:`~opacus.optimizers.optimizer.DPOptimizer` that implements
-    per layer clipping strategy and is compatible with distibured data parallel
+    per layer clipping strategy and is compatible with distributed data parallel
     """
 
     def __init__(

opacus/optimizers/optimizer.py

@@ -113,7 +113,7 @@ def _generate_noise(
         reference: The reference Tensor to get the appropriate shape and device
             for generating the noise
         generator: The PyTorch noise generator
-        secure_mode: boolean showing if "secure" noise need to be generate
+        secure_mode: boolean showing if "secure" noise need to be generated
             (see the notes)
 
     Notes:
@@ -186,7 +186,7 @@ class DPOptimizer(Optimizer):
     Examples:
         >>> module = MyCustomModel()
         >>> optimizer = torch.optim.SGD(module.parameters(), lr=0.1)
-        >>> dp_optimzer = DPOptimizer(
+        >>> dp_optimizer = DPOptimizer(
         ...     optimizer=optimizer,
         ...     noise_multiplier=1.0,
         ...     max_grad_norm=1.0,

opacus/utils/module_utils.py

@@ -72,7 +72,7 @@ def requires_grad(module: nn.Module, *, recurse: bool = False) -> bool:
     Args:
         module: PyTorch module whose parameters are to be examined.
         recurse: Flag specifying if the gradient requirement check should
-            be applied recursively to sub-modules of the specified module
+            be applied recursively to submodules of the specified module
 
     Returns:
         Flag indicate if any parameters require gradients

opacus/utils/tensor_utils.py

@@ -30,7 +30,7 @@ def calc_sample_norms(
     Calculates the norm of the given tensors for each sample.
 
     This function calculates the overall norm of the given tensors for each sample,
-    assuming the each batch's dim is zero.
+    assuming each batch's dim is zero.
 
     Args:
         named_params: An iterator of tuples <name, param> with name being a
@@ -61,7 +61,7 @@ def calc_sample_norms_one_layer(param: torch.Tensor) -> torch.Tensor:
     Calculates the norm of the given tensor (a single parameter) for each sample.
 
     This function calculates the overall norm of the given tensor for each sample,
-    assuming the each batch's dim is zero.
+    assuming each batch's dim is zero.
 
     It is equivalent to:
     `calc_sample_norms(named_params=((None, param),))[0]`
@@ -90,7 +90,7 @@ def sum_over_all_but_batch_and_last_n(
     Calculates the sum over all dimensions, except the first
     (batch dimension), and excluding the last n_dims.
 
-    This function will ignore the first dimension and it will
+    This function will ignore the first dimension, and it will
     not aggregate over the last n_dims dimensions.
 
     Args:

opacus/utils/uniform_sampler.py

@@ -68,7 +68,7 @@ class DistributedUniformWithReplacementSampler(Sampler):
            (plus or minus one sample)
         3. Each replica selects each sample of its chunk independently
            with probability `sample_rate`
-        4. Each replica ouputs the selected samples, which form a local batch
+        4. Each replica outputs the selected samples, which form a local batch
 
     The sum of the lengths of the local batches follows a Poisson distribution.
     In particular, the expected length of each local batch is: