Add many missing spaces in adjacent strings (#26751)

Add missing spaces in adjacent strings

examples/pytorch/image-classification/run_image_classification_no_trainer.py

@@ -152,7 +152,7 @@ def parse_args():
         default=False,
         help=(
             "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-            "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+            "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
             "execute code present on the Hub on your local machine."
         ),
     )
@@ -179,7 +179,7 @@ def parse_args():
         default="all",
         help=(
             'The integration to report the results and logs to. Supported platforms are `"tensorboard"`,'
-            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.'
+            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations. '
             "Only applicable when `--with_tracking` is passed."
         ),
     )

examples/pytorch/image-pretraining/run_mim.py

@@ -174,7 +174,7 @@ class ModelArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },

examples/pytorch/image-pretraining/run_mim_no_trainer.py

@@ -208,7 +208,7 @@ def parse_args():
         default=False,
         help=(
             "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-            "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+            "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
             "execute code present on the Hub on your local machine."
         ),
     )
@@ -246,7 +246,7 @@ def parse_args():
         default="all",
         help=(
             'The integration to report the results and logs to. Supported platforms are `"tensorboard"`,'
-            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.'
+            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations. '
             "Only applicable when `--with_tracking` is passed."
         ),
     )

examples/pytorch/language-modeling/run_clm.py

@@ -132,7 +132,7 @@ class ModelArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },
@@ -151,7 +151,7 @@ class ModelArguments:
         default=False,
         metadata={
             "help": (
-                "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded."
+                "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded. "
                 "set True will benefit LLM loading time and RAM consumption."
             )
         },
@@ -424,7 +424,7 @@ def main():
         tokenizer = AutoTokenizer.from_pretrained(model_args.model_name_or_path, **tokenizer_kwargs)
     else:
         raise ValueError(
-            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script. "
             "You can do it from another script, save it, and load it from here, using --tokenizer_name."
         )
 
@@ -506,7 +506,7 @@ def main():
     else:
         if data_args.block_size > tokenizer.model_max_length:
             logger.warning(
-                f"The block_size passed ({data_args.block_size}) is larger than the maximum length for the model"
+                f"The block_size passed ({data_args.block_size}) is larger than the maximum length for the model "
                 f"({tokenizer.model_max_length}). Using block_size={tokenizer.model_max_length}."
             )
         block_size = min(data_args.block_size, tokenizer.model_max_length)

examples/pytorch/language-modeling/run_clm_no_trainer.py

@@ -199,7 +199,7 @@ def parse_args():
         default=False,
         help=(
             "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-            "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+            "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
             "execute code present on the Hub on your local machine."
         ),
     )
@@ -226,7 +226,7 @@ def parse_args():
         default="all",
         help=(
             'The integration to report the results and logs to. Supported platforms are `"tensorboard"`,'
-            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.'
+            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations. '
             "Only applicable when `--with_tracking` is passed."
         ),
     )
@@ -234,7 +234,7 @@ def parse_args():
         "--low_cpu_mem_usage",
         action="store_true",
         help=(
-            "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded."
+            "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded. "
             "If passed, LLM loading time and RAM consumption will be benefited."
         ),
     )
@@ -398,7 +398,7 @@ def main():
         )
     else:
         raise ValueError(
-            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script. "
             "You can do it from another script, save it, and load it from here, using --tokenizer_name."
         )
 
@@ -449,7 +449,7 @@ def main():
     else:
         if args.block_size > tokenizer.model_max_length:
             logger.warning(
-                f"The block_size passed ({args.block_size}) is larger than the maximum length for the model"
+                f"The block_size passed ({args.block_size}) is larger than the maximum length for the model "
                 f"({tokenizer.model_max_length}). Using block_size={tokenizer.model_max_length}."
             )
         block_size = min(args.block_size, tokenizer.model_max_length)

examples/pytorch/language-modeling/run_mlm.py

@@ -128,7 +128,7 @@ class ModelArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },
@@ -137,7 +137,7 @@ class ModelArguments:
         default=False,
         metadata={
             "help": (
-                "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded."
+                "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded. "
                 "set True will benefit LLM loading time and RAM consumption."
             )
         },
@@ -417,7 +417,7 @@ def main():
         tokenizer = AutoTokenizer.from_pretrained(model_args.model_name_or_path, **tokenizer_kwargs)
     else:
         raise ValueError(
-            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script. "
             "You can do it from another script, save it, and load it from here, using --tokenizer_name."
         )
 
@@ -462,7 +462,7 @@ def main():
     else:
         if data_args.max_seq_length > tokenizer.model_max_length:
             logger.warning(
-                f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
+                f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the "
                 f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
             )
         max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)

examples/pytorch/language-modeling/run_mlm_no_trainer.py

@@ -206,7 +206,7 @@ def parse_args():
         default=False,
         help=(
             "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-            "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+            "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
             "execute code present on the Hub on your local machine."
         ),
     )
@@ -233,7 +233,7 @@ def parse_args():
         default="all",
         help=(
             'The integration to report the results and logs to. Supported platforms are `"tensorboard"`,'
-            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.'
+            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations. '
             "Only applicable when `--with_tracking` is passed."
         ),
     )
@@ -241,7 +241,7 @@ def parse_args():
         "--low_cpu_mem_usage",
         action="store_true",
         help=(
-            "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded."
+            "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded. "
             "If passed, LLM loading time and RAM consumption will be benefited."
         ),
     )
@@ -395,7 +395,7 @@ def main():
         )
     else:
         raise ValueError(
-            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script. "
             "You can do it from another script, save it, and load it from here, using --tokenizer_name."
         )
 
@@ -434,7 +434,7 @@ def main():
     else:
         if args.max_seq_length > tokenizer.model_max_length:
             logger.warning(
-                f"The max_seq_length passed ({args.max_seq_length}) is larger than the maximum length for the"
+                f"The max_seq_length passed ({args.max_seq_length}) is larger than the maximum length for the "
                 f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
             )
         max_seq_length = min(args.max_seq_length, tokenizer.model_max_length)

examples/pytorch/language-modeling/run_plm.py

@@ -115,7 +115,7 @@ class ModelArguments:
         default=False,
         metadata={
             "help": (
-                "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded."
+                "It is an option to create the model as an empty shell, then only materialize its parameters when the pretrained weights are loaded. "
                 "set True will benefit LLM loading time and RAM consumption."
             )
         },
@@ -385,7 +385,7 @@ def main():
         tokenizer = AutoTokenizer.from_pretrained(model_args.model_name_or_path, **tokenizer_kwargs)
     else:
         raise ValueError(
-            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script. "
             "You can do it from another script, save it, and load it from here, using --tokenizer_name."
         )
 
@@ -419,7 +419,7 @@ def main():
 
     if data_args.max_seq_length > tokenizer.model_max_length:
         logger.warning(
-            f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
+            f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the "
             f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
         )
     max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)

examples/pytorch/multiple-choice/run_swag.py

@@ -100,7 +100,7 @@ class ModelArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },
@@ -375,7 +375,7 @@ def main():
     else:
         if data_args.max_seq_length > tokenizer.model_max_length:
             logger.warning(
-                f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
+                f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the "
                 f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
             )
         max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)

examples/pytorch/multiple-choice/run_swag_no_trainer.py

@@ -188,7 +188,7 @@ def parse_args():
         default=False,
         help=(
             "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-            "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+            "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
             "execute code present on the Hub on your local machine."
         ),
     )
@@ -215,7 +215,7 @@ def parse_args():
         default="all",
         help=(
             'The integration to report the results and logs to. Supported platforms are `"tensorboard"`,'
-            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.'
+            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations. '
             "Only applicable when `--with_tracking` is passed."
         ),
     )
@@ -401,7 +401,7 @@ def main():
         )
     else:
         raise ValueError(
-            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script. "
             "You can do it from another script, save it, and load it from here, using --tokenizer_name."
         )
 

examples/pytorch/question-answering/run_qa.py

@@ -100,7 +100,7 @@ class ModelArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },
@@ -391,7 +391,7 @@ def main():
 
     if data_args.max_seq_length > tokenizer.model_max_length:
         logger.warning(
-            f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
+            f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the "
             f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
         )
     max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)

examples/pytorch/question-answering/run_qa_beam_search.py

@@ -367,7 +367,7 @@ def main():
 
     if data_args.max_seq_length > tokenizer.model_max_length:
         logger.warning(
-            f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
+            f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the "
             f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
         )
     max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)

examples/pytorch/question-answering/run_qa_beam_search_no_trainer.py

@@ -395,7 +395,7 @@ def main():
 
     if args.max_seq_length > tokenizer.model_max_length:
         logger.warning(
-            f"The max_seq_length passed ({args.max_seq_length}) is larger than the maximum length for the"
+            f"The max_seq_length passed ({args.max_seq_length}) is larger than the maximum length for the "
             f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
         )
 

examples/pytorch/question-answering/run_qa_no_trainer.py

@@ -279,7 +279,7 @@ def parse_args():
         default=False,
         help=(
             "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-            "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+            "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
             "execute code present on the Hub on your local machine."
         ),
     )
@@ -306,7 +306,7 @@ def parse_args():
         default="all",
         help=(
             'The integration to report the results and logs to. Supported platforms are `"tensorboard"`,'
-            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.'
+            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations. '
             "Only applicable when `--with_tracking` is passed."
         ),
     )
@@ -442,7 +442,7 @@ def main():
         )
     else:
         raise ValueError(
-            "You are instantiating a new tokenizer from scratch. This is not supported by this script."
+            "You are instantiating a new tokenizer from scratch. This is not supported by this script. "
             "You can do it from another script, save it, and load it from here, using --tokenizer_name."
         )
 
@@ -471,7 +471,7 @@ def main():
 
     if args.max_seq_length > tokenizer.model_max_length:
         logger.warning(
-            f"The max_seq_length passed ({args.max_seq_length}) is larger than the maximum length for the"
+            f"The max_seq_length passed ({args.max_seq_length}) is larger than the maximum length for the "
             f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
         )
 

examples/pytorch/question-answering/run_seq2seq_qa.py

@@ -101,7 +101,7 @@ class ModelArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },
@@ -171,7 +171,7 @@ class DataTrainingArguments:
         metadata={
             "help": (
                 "The maximum total sequence length for validation target text after tokenization. Sequences longer "
-                "than this will be truncated, sequences shorter will be padded. Will default to `max_answer_length`."
+                "than this will be truncated, sequences shorter will be padded. Will default to `max_answer_length`. "
                 "This argument is also used to override the ``max_length`` param of ``model.generate``, which is used "
                 "during ``evaluate`` and ``predict``."
             )
@@ -465,13 +465,13 @@ def main():
 
     if training_args.label_smoothing_factor > 0 and not hasattr(model, "prepare_decoder_input_ids_from_labels"):
         logger.warning(
-            "label_smoothing is enabled but the `prepare_decoder_input_ids_from_labels` method is not defined for"
+            "label_smoothing is enabled but the `prepare_decoder_input_ids_from_labels` method is not defined for "
             f"`{model.__class__.__name__}`. This will lead to loss being calculated twice and will take up more memory"
         )
 
     if data_args.max_seq_length > tokenizer.model_max_length:
         logger.warning(
-            f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
+            f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the "
             f"model ({tokenizer.model_max_length}). Using max_seq_length={tokenizer.model_max_length}."
         )
     max_seq_length = min(data_args.max_seq_length, tokenizer.model_max_length)

examples/pytorch/semantic-segmentation/run_semantic_segmentation.py

@@ -262,7 +262,7 @@ class ModelArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },

examples/pytorch/semantic-segmentation/run_semantic_segmentation_no_trainer.py

@@ -279,7 +279,7 @@ def parse_args():
         default=False,
         help=(
             "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-            "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+            "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
             "execute code present on the Hub on your local machine."
         ),
     )
@@ -307,7 +307,7 @@ def parse_args():
         default="all",
         help=(
             'The integration to report the results and logs to. Supported platforms are `"tensorboard"`,'
-            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.'
+            ' `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations. '
             "Only applicable when `--with_tracking` is passed."
         ),
     )

examples/pytorch/speech-recognition/run_speech_recognition_ctc.py

@@ -104,8 +104,8 @@ class ModelArguments:
         default=0.05,
         metadata={
             "help": (
-                "Probability of each feature vector along the time axis to be chosen as the start of the vector"
-                "span to be masked. Approximately ``mask_time_prob * sequence_length // mask_time_length`` feature"
+                "Probability of each feature vector along the time axis to be chosen as the start of the vector "
+                "span to be masked. Approximately ``mask_time_prob * sequence_length // mask_time_length`` feature "
                 "vectors will be masked along the time axis."
             )
         },
@@ -249,7 +249,7 @@ class DataTrainingArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },
@@ -430,7 +430,7 @@ def main():
 
     # Log on each process the small summary:
     logger.warning(
-        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
+        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu} "
         f"distributed training: {training_args.parallel_mode.value == 'distributed'}, 16-bits training: {training_args.fp16}"
     )
     # Set the verbosity to info of the Transformers logger (on main process only):

examples/pytorch/speech-recognition/run_speech_recognition_ctc_adapter.py

@@ -90,8 +90,8 @@ class ModelArguments:
         default=0.05,
         metadata={
             "help": (
-                "Probability of each feature vector along the time axis to be chosen as the start of the vector"
-                "span to be masked. Approximately ``mask_time_prob * sequence_length // mask_time_length`` feature"
+                "Probability of each feature vector along the time axis to be chosen as the start of the vector "
+                "span to be masked. Approximately ``mask_time_prob * sequence_length // mask_time_length`` feature "
                 "vectors will be masked along the time axis."
             )
         },
@@ -252,7 +252,7 @@ class DataTrainingArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },
@@ -426,7 +426,7 @@ def main():
 
     # Log on each process the small summary:
     logger.warning(
-        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
+        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu} "
         f"distributed training: {training_args.parallel_mode.value == 'distributed'}, 16-bits training: {training_args.fp16}"
     )
     # Set the verbosity to info of the Transformers logger (on main process only):

examples/pytorch/speech-recognition/run_speech_recognition_seq2seq.py

@@ -106,7 +106,7 @@ class ModelArguments:
         metadata={
             "help": (
                 "Whether or not to allow for custom models defined on the Hub in their own modeling files. This option"
-                "should only be set to `True` for repositories you trust and in which you have read the code, as it will"
+                "should only be set to `True` for repositories you trust and in which you have read the code, as it will "
                 "execute code present on the Hub on your local machine."
             )
         },
@@ -322,7 +322,7 @@ def main():
 
     # Log on each process the small summary:
     logger.warning(
-        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
+        f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu} "
         f"distributed training: {training_args.parallel_mode.value == 'distributed'}, 16-bits training: {training_args.fp16}"
     )
     logger.info(f"Training/evaluation parameters {training_args}")