Fixed minor spelling mistakes (#10489)

Co-authored-by: WybeKoper <WybeKoper@users.noreply.github.com>

src/transformers/data/processors/squad.py

@@ -779,7 +779,7 @@ class SquadFeatures:
         token_to_orig_map: mapping between the tokens and the original text, needed in order to identify the answer.
         start_position: start of the answer token index
         end_position: end of the answer token index
-        encoding: optionally store the BatchEncoding with the fast-tokenizer alignement methods.
+        encoding: optionally store the BatchEncoding with the fast-tokenizer alignment methods.
     """
 
     def __init__(

src/transformers/models/ibert/quant_modules.py

@@ -169,7 +169,7 @@ def forward(
     ):
 
         x_act = x if identity is None else identity + x
-        # collect runnng stats if traiing
+        # collect running stats if training
         if self.training:
             assert not self.percentile, "percentile mode is not currently supported for activation."
             assert not self.per_channel, "per-channel mode is not currently supported for activation."
@@ -746,7 +746,7 @@ def batch_frexp(inputs, max_bit=31):
 
 class FixedPointMul(Function):
     """
-    Function to perform fixed-point arthmetic that can match integer arthmetic on hardware.
+    Function to perform fixed-point arithmetic that can match integer arithmetic on hardware.
 
     Args:
         pre_act (:obj:`torch.Tensor`):

src/transformers/trainer_pt_utils.py

@@ -44,7 +44,7 @@
 if is_torch_tpu_available():
     import torch_xla.core.xla_model as xm
 
-# this is used to supress an undesired warning emitted by pytorch versions 1.4.2-1.7.0
+# this is used to suppress an undesired warning emitted by pytorch versions 1.4.2-1.7.0
 try:
     from torch.optim.lr_scheduler import SAVE_STATE_WARNING
 except ImportError: