diff --git a/examples/question-answering/run_qa.py b/examples/question-answering/run_qa.py
index 064717d80f..a590986d58 100644
--- a/examples/question-answering/run_qa.py
+++ b/examples/question-answering/run_qa.py
@@ -378,8 +378,8 @@ def main():
     model = model.to("hpu")
     # Tokenizer check: this script requires a fast tokenizer.
     if not isinstance(tokenizer, PreTrainedTokenizerFast):
-        raise ValueError(
-            "This example script only works for models that have a fast tokenizer. Checkout the big table of models at"
+        raise TypeError(
+            "This example script only works for models that have a fast tokenizer. Check out the big table of models at"
             " https://huggingface.co/transformers/index.html#supported-frameworks to find the model types that meet"
             " this requirement"
         )
diff --git a/optimum/habana/transformers/generation/utils.py b/optimum/habana/transformers/generation/utils.py
index a8b1858e99..ed33b99534 100755
--- a/optimum/habana/transformers/generation/utils.py
+++ b/optimum/habana/transformers/generation/utils.py
@@ -2998,7 +2998,7 @@ def _beam_search(
                 f"Batch dimension of `input_ids` should be {num_beams * batch_size}, but is {batch_beam_size}."
             )
 
-        # (joao) feature lost in the refactor. Probably won't implement, hurts readbility with minimal gains (there
+        # (joao) feature lost in the refactor. Probably won't implement, hurts readability with minimal gains (there
         # are newer low-memory alternatives like the offloaded cache)
         sequential = generation_config.low_memory
         if sequential:
diff --git a/optimum/habana/transformers/loss/loss_rt_detr.py b/optimum/habana/transformers/loss/loss_rt_detr.py
index 50cdc3a3b7..b119cd8de6 100644
--- a/optimum/habana/transformers/loss/loss_rt_detr.py
+++ b/optimum/habana/transformers/loss/loss_rt_detr.py
@@ -54,7 +54,7 @@ def gaudi_RTDetrHungarianMatcher_forward(self, outputs, targets):
     target_bbox = torch.cat([v["boxes"] for v in targets])
     # Compute the classification cost. Contrary to the loss, we don't use the NLL,
     # but approximate it in 1 - proba[target class].
-    # The 1 is a constant that doesn't change the matching, it can be ommitted.
+    # The 1 is a constant that doesn't change the matching, it can be omitted.
     if self.use_focal_loss:
         out_prob = F.sigmoid(outputs["logits"].flatten(0, 1))
         out_prob = out_prob[:, target_ids]
@@ -67,7 +67,7 @@ def gaudi_RTDetrHungarianMatcher_forward(self, outputs, targets):
 
     # Compute the L1 cost between boxes
     bbox_cost = torch.cdist(out_bbox, target_bbox, p=1)
-    # Compute the giou cost betwen boxes
+    # Compute the giou cost between boxes
     giou_cost = -generalized_box_iou(center_to_corners_format(out_bbox), center_to_corners_format(target_bbox))
     # Compute the final cost matrix
     cost_matrix = self.bbox_cost * bbox_cost + self.class_cost * class_cost + self.giou_cost * giou_cost