Fix typos in strings and comments (#37910)

examples/legacy/multiple_choice/utils_multiple_choice.py

@@ -539,7 +539,7 @@ def convert_examples_to_features(
             if "num_truncated_tokens" in inputs and inputs["num_truncated_tokens"] > 0:
                 logger.info(
                     "Attention! you are cropping tokens (swag task is ok). "
-                    "If you are training ARC and RACE and you are poping question + options, "
+                    "If you are training ARC and RACE and you are popping question + options, "
                     "you need to try to use a bigger max seq length!"
                 )
 

examples/legacy/question-answering/run_squad.py

@@ -745,7 +745,7 @@ def main():
         args.tokenizer_name if args.tokenizer_name else args.model_name_or_path,
         do_lower_case=args.do_lower_case,
         cache_dir=args.cache_dir if args.cache_dir else None,
-        use_fast=False,  # SquadDataset is not compatible with Fast tokenizers which have a smarter overflow handeling
+        use_fast=False,  # SquadDataset is not compatible with Fast tokenizers which have a smarter overflow handling
     )
     model = AutoModelForQuestionAnswering.from_pretrained(
         args.model_name_or_path,
@@ -795,7 +795,7 @@ def main():
         # Load a trained model and vocabulary that you have fine-tuned
         model = AutoModelForQuestionAnswering.from_pretrained(args.output_dir)  # , force_download=True)
 
-        # SquadDataset is not compatible with Fast tokenizers which have a smarter overflow handeling
+        # SquadDataset is not compatible with Fast tokenizers which have a smarter overflow handling
         # So we use use_fast=False here for now until Fast-tokenizer-compatible-examples are out
         tokenizer = AutoTokenizer.from_pretrained(args.output_dir, do_lower_case=args.do_lower_case, use_fast=False)
         model.to(args.device)

examples/legacy/question-answering/run_squad_trainer.py

@@ -122,7 +122,7 @@ def main():
     tokenizer = AutoTokenizer.from_pretrained(
         model_args.tokenizer_name if model_args.tokenizer_name else model_args.model_name_or_path,
         cache_dir=model_args.cache_dir,
-        use_fast=False,  # SquadDataset is not compatible with Fast tokenizers which have a smarter overflow handeling
+        use_fast=False,  # SquadDataset is not compatible with Fast tokenizers which have a smarter overflow handling
     )
     model = AutoModelForQuestionAnswering.from_pretrained(
         model_args.model_name_or_path,

examples/legacy/run_transfo_xl.py

@@ -71,7 +71,7 @@ def main():
     # You can also build the corpus yourself using TransfoXLCorpus methods
     # The pre-processing involve computing word frequencies to prepare the Adaptive input and SoftMax
     # and tokenizing the dataset
-    # The pre-processed corpus is a convertion (using the conversion script )
+    # The pre-processed corpus is a conversion (using the conversion script )
     corpus = TransfoXLCorpus.from_pretrained(args.model_name)
 
     va_iter = corpus.get_iterator("valid", args.batch_size, args.tgt_len, device=device, ext_len=args.ext_len)

examples/legacy/seq2seq/pack_dataset.py

@@ -40,7 +40,7 @@ def pack_examples(tok, src_examples, tgt_examples, max_tokens=1024):
     for src, tgt in tqdm(sorted_examples[1:]):
         cand_src = new_src + " " + src
         cand_tgt = new_tgt + " " + tgt
-        if is_too_big(cand_src) or is_too_big(cand_tgt):  # cant fit, finalize example
+        if is_too_big(cand_src) or is_too_big(cand_tgt):  # can't fit, finalize example
             finished_src.append(new_src)
             finished_tgt.append(new_tgt)
             new_src, new_tgt = src, tgt

examples/pytorch/speech-recognition/run_speech_recognition_ctc_adapter.py

@@ -804,7 +804,7 @@ def main():
     if "common_voice" in data_args.dataset_name:
         kwargs["language"] = config_name
 
-    # make sure that adapter weights are saved seperately
+    # make sure that adapter weights are saved separately
     adapter_file = WAV2VEC2_ADAPTER_SAFE_FILE.format(data_args.target_language)
     adapter_file = os.path.join(training_args.output_dir, adapter_file)
     logger.info(f"Saving adapter weights under {adapter_file}...")

src/transformers/integrations/executorch.py

@@ -516,7 +516,7 @@ def convert_and_export_with_cache(
                     "Dynamic shapes spec will be ignored by convert_and_export_with_cache for torch < 2.6.0."
                 )
             if strict is not None:
-                logging.warning("The strict flag will be ingored by convert_and_export_with_cache for torch < 2.6.0.")
+                logging.warning("The strict flag will be ignored by convert_and_export_with_cache for torch < 2.6.0.")
             # We have to keep this path for BC.
             #
             # Due to issue https://github.com/pytorch/pytorch/issues/128394, we need to switch to use an internal

src/transformers/loss/loss_d_fine.py

@@ -152,8 +152,8 @@ def bbox2distance(points, bbox, max_num_bins, reg_scale, up, eps=0.1):
         points (Tensor): (n, 4) [x, y, w, h], where (x, y) is the center.
         bbox (Tensor): (n, 4) bounding boxes in "xyxy" format.
         max_num_bins (float): Maximum bin value.
-        reg_scale (float): Controling curvarture of W(n).
+        reg_scale (float): Controlling curvarture of W(n).
-        up (Tensor): Controling upper bounds of W(n).
+        up (Tensor): Controlling upper bounds of W(n).
         eps (float): Small value to ensure target < max_num_bins.
 
     Returns:

src/transformers/models/d_fine/configuration_d_fine.py

@@ -28,7 +28,7 @@ logger = logging.get_logger(__name__)
 
 
 # TODO: Attribute map assignment logic should be fixed in modular
-# as well as super() call parsing becuase otherwise we cannot re-write args after initialization
+# as well as super() call parsing because otherwise we cannot re-write args after initialization
 class DFineConfig(PretrainedConfig):
     """
     This is the configuration class to store the configuration of a [`DFineModel`]. It is used to instantiate a D-FINE

src/transformers/models/d_fine/modular_d_fine.py

@@ -47,7 +47,7 @@ logger = logging.get_logger(__name__)
 
 
 # TODO: Attribute map assignment logic should be fixed in modular
-# as well as super() call parsing becuase otherwise we cannot re-write args after initialization
+# as well as super() call parsing because otherwise we cannot re-write args after initialization
 class DFineConfig(PretrainedConfig):
     """
     This is the configuration class to store the configuration of a [`DFineModel`]. It is used to instantiate a D-FINE

src/transformers/models/sam_hq/modeling_sam_hq.py

@@ -1615,7 +1615,7 @@ class SamHQModel(SamHQPreTrainedModel):
                 Input boxes for the points, this is used by the prompt encoder to encode the prompt. Generally yields to
                 much better generated masks. The boxes can be obtained by passing a list of list of list to the processor,
                 that will generate a `torch` tensor, with each dimension corresponding respectively to the image batch
-                size, the number of boxes per image and the coordinates of the top left and botton right point of the box.
+                size, the number of boxes per image and the coordinates of the top left and bottom right point of the box.
                 In the order (`x1`, `y1`, `x2`, `y2`):
 
                 - `x1`: the x coordinate of the top left point of the input box

src/transformers/models/sam_hq/modular_sam_hq.py

@@ -551,7 +551,7 @@ class SamHQModel(SamModel):
                 Input boxes for the points, this is used by the prompt encoder to encode the prompt. Generally yields to
                 much better generated masks. The boxes can be obtained by passing a list of list of list to the processor,
                 that will generate a `torch` tensor, with each dimension corresponding respectively to the image batch
-                size, the number of boxes per image and the coordinates of the top left and botton right point of the box.
+                size, the number of boxes per image and the coordinates of the top left and bottom right point of the box.
                 In the order (`x1`, `y1`, `x2`, `y2`):
 
                 - `x1`: the x coordinate of the top left point of the input box

tests/models/d_fine/test_modeling_d_fine.py

@@ -586,10 +586,10 @@ class DFineModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
                     self.model_tester.num_labels,
                 )
                 self.assertEqual(outputs.logits.shape, expected_shape)
-                # Confirm out_indices was propogated to backbone
+                # Confirm out_indices was propagated to backbone
                 self.assertEqual(len(model.model.backbone.intermediate_channel_sizes), 3)
             else:
-                # Confirm out_indices was propogated to backbone
+                # Confirm out_indices was propagated to backbone
                 self.assertEqual(len(model.backbone.intermediate_channel_sizes), 3)
 
             self.assertTrue(outputs)
@@ -618,10 +618,10 @@ class DFineModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
                     self.model_tester.num_labels,
                 )
                 self.assertEqual(outputs.logits.shape, expected_shape)
-                # Confirm out_indices was propogated to backbone
+                # Confirm out_indices was propagated to backbone
                 self.assertEqual(len(model.model.backbone.intermediate_channel_sizes), 3)
             else:
-                # Confirm out_indices was propogated to backbone
+                # Confirm out_indices was propagated to backbone
                 self.assertEqual(len(model.backbone.intermediate_channel_sizes), 3)
 
             self.assertTrue(outputs)

tests/models/fsmt/test_modeling_fsmt.py

@@ -423,7 +423,7 @@ class FSMTHeadTests(unittest.TestCase):
 
 
 def _assert_tensors_equal(a, b, atol=1e-12, prefix=""):
-    """If tensors not close, or a and b arent both tensors, raise a nice Assertion error."""
+    """If tensors not close, or a and b aren't both tensors, raise a nice Assertion error."""
     if a is None and b is None:
         return True
     try:

tests/models/gemma2/test_modeling_gemma2.py

@@ -149,7 +149,7 @@ class Gemma2ModelTest(GemmaModelTest, unittest.TestCase):
 
     @unittest.skip(
         reason="HybridCache can't be gathered because it is not iterable. Adding a simple iter and dumping `distributed_iterator`"
-        " as in Dynamic Cache doesnt work. NOTE: @gante all cache objects would need better compatibility with multi gpu setting"
+        " as in Dynamic Cache doesn't work. NOTE: @gante all cache objects would need better compatibility with multi gpu setting"
     )
     def test_multi_gpu_data_parallel_forward(self):
         pass

tests/models/qwen2_5_omni/test_processor_qwen2_5_omni.py

@@ -567,7 +567,7 @@ class Qwen2_5OmniProcessorTest(ProcessorTesterMixin, unittest.TestCase):
             signature.parameters.get("videos") is not None
             and signature.parameters["videos"].annotation == inspect._empty
         ):
-            self.skipTest(f"{self.processor_class} does not suport video inputs")
+            self.skipTest(f"{self.processor_class} does not support video inputs")
 
         if "feature_extractor" not in self.processor_class.attributes:
             self.skipTest(f"feature_extractor attribute not present in {self.processor_class}")

tests/models/sam_hq/test_modeling_sam_hq.py

@@ -244,13 +244,13 @@ class SamHQVisionModelTest(ModelTesterMixin, unittest.TestCase):
         pass
 
     @unittest.skip(
-        reason="This architecure seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
+        reason="This architecture seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
     )
     def test_training_gradient_checkpointing_use_reentrant(self):
         pass
 
     @unittest.skip(
-        reason="This architecure seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
+        reason="This architecture seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
     )
     def test_training_gradient_checkpointing_use_reentrant_false(self):
         pass
@@ -682,13 +682,13 @@ class SamHQModelTest(ModelTesterMixin, PipelineTesterMixin, unittest.TestCase):
         pass
 
     @unittest.skip(
-        reason="This architecure seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
+        reason="This architecture seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
     )
     def test_training_gradient_checkpointing_use_reentrant(self):
         pass
 
     @unittest.skip(
-        reason="This architecure seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
+        reason="This architecture seem to not compute gradients properly when using GC, check: https://github.com/huggingface/transformers/pull/27124"
     )
     def test_training_gradient_checkpointing_use_reentrant_false(self):
         pass