Fix typos

docs/source/de/quicktour.md

@@ -156,7 +156,7 @@ Die [`pipeline`] kann jedes Modell aus dem [Model Hub](https://huggingface.co/mo
 
 <frameworkcontent>
 <pt>
-Use the [`AutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the pretrained model and it's associated tokenizer (more on an `AutoClass` below):
+Use the [`AutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the pretrained model and its associated tokenizer (more on an `AutoClass` below):
 
 ```py
 >>> from transformers import AutoTokenizer, AutoModelForSequenceClassification
@@ -166,7 +166,7 @@ Use the [`AutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the
 ```
 </pt>
 <tf>
-Use the [`TFAutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the pretrained model and it's associated tokenizer (more on an `TFAutoClass` below):
+Use the [`TFAutoModelForSequenceClassification`] and [`AutoTokenizer`] to load the pretrained model and its associated tokenizer (more on an `TFAutoClass` below):
 
 ```py
 >>> from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
@@ -222,7 +222,7 @@ Anschließend wandelt der Tokenizer die Token in Zahlen um, um einen Tensor als
 Der Tokenizer gibt ein Wörterbuch zurück, das Folgendes enthält:
 
 * [input_ids](./glossary#input-ids): numerische Repräsentationen Ihrer Token.
-* [atttention_mask](.glossary#attention-mask): gibt an, welche Token beachtet werden sollen.
+* [attention_mask](.glossary#attention-mask): gibt an, welche Token beachtet werden sollen.
 
 Genau wie die [`pipeline`] akzeptiert der Tokenizer eine Liste von Eingaben. Darüber hinaus kann der Tokenizer den Text auch auffüllen und kürzen, um einen Stapel mit einheitlicher Länge zurückzugeben:
 

docs/source/en/cache_explanation.md

@@ -9,7 +9,7 @@ Unless required by applicable law or agreed to in writing, software distributed
 an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
 specific language governing permissions and limitations under the License.
 
-⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
+⚠️ Note that this file is in Markdown but contains specific syntax for our doc-builder (similar to MDX) that may not be
 rendered properly in your Markdown viewer.
 
 -->
@@ -62,7 +62,7 @@ for _ in range(max_new_tokens):
     # Greedily sample one next token
     next_token_ids = outputs.logits[:, -1:].argmax(-1)
     generated_ids = torch.cat([generated_ids, next_token_ids], dim=-1)
-    # Prepare inputs for the next generation step by leaaving unprocessed tokens, in our case we have only one new token
+    # Prepare inputs for the next generation step by leaving unprocessed tokens, in our case we have only one new token
     # and expanding attn mask for the new token, as explained above
     attention_mask = inputs["attention_mask"]
     attention_mask = torch.cat([attention_mask, attention_mask.new_ones((attention_mask.shape[0], 1))], dim=-1)
@@ -88,7 +88,7 @@ model = AutoModelForCausalLM.from_pretrained("meta-llama/Llama-2-7b-chat-hf", to
 inputs = tokenizer("Hello, my name is", return_tensors="pt").to(model.device)
 
 # `return_dict_in_generate=True` is required to return the cache and `return_legacy_cache` forces the returned cache
-# in the the legacy format
+# in the legacy format
 generation_outputs = model.generate(**inputs, return_dict_in_generate=True, return_legacy_cache=True, max_new_tokens=5)
 
 cache = DynamicCache.from_legacy_cache(generation_outputs.past_key_values)

docs/source/en/chat_templating_multimodal.md

@@ -9,7 +9,7 @@ Unless required by applicable law or agreed to in writing, software distributed
 an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
 specific language governing permissions and limitations under the License.
 
-⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
+⚠️ Note that this file is in Markdown but contains specific syntax for our doc-builder (similar to MDX) that may not be
 rendered properly in your Markdown viewer.
 
 -->
@@ -18,7 +18,7 @@ rendered properly in your Markdown viewer.
 
 Multimodal model chat templates expect a similar [template](./chat_templating) as text-only models. It needs `messages` that includes a dictionary of the `role` and `content`.
 
-Multimodal templates are included in the [Processor](./processors) class and requires an additional `type` key for specifying whether the included content is an image, video, or text.
+Multimodal templates are included in the [Processor](./processors) class and require an additional `type` key for specifying whether the included content is an image, video, or text.
 
 This guide will show you how to format chat templates for multimodal models as well as some best practices for configuring the template
 
@@ -109,7 +109,7 @@ These inputs are now ready to be used in [`~GenerationMixin.generate`].
 
 Some vision models also support video inputs. The message format is very similar to the format for [image inputs](#image-inputs).
 
-- The content `"type"` should be `"video"` to indicate the the content is a video.
+- The content `"type"` should be `"video"` to indicate the content is a video.
 - For videos, it can be a link to the video (`"url"`) or it could be a file path (`"path"`). Videos loaded from a URL can only be decoded with [PyAV](https://pyav.basswood-io.com/docs/stable/) or [Decord](https://github.com/dmlc/decord).
 
 > [!WARNING]
@@ -141,7 +141,7 @@ Pass `messages` to [`~ProcessorMixin.apply_chat_template`] to tokenize the input
 
 The `video_load_backend` parameter refers to a specific framework to load a video. It supports [PyAV](https://pyav.basswood-io.com/docs/stable/), [Decord](https://github.com/dmlc/decord), [OpenCV](https://github.com/opencv/opencv), and [torchvision](https://pytorch.org/vision/stable/index.html).
 
-The examples below uses Decord as the backend because it is a bit faster than PyAV.
+The examples below use Decord as the backend because it is a bit faster than PyAV.
 
 <hfoptions id="sampling">
 <hfoption id="fixed number of frames">

docs/source/en/custom_models.md

@@ -131,7 +131,7 @@ class ResnetModel(PreTrainedModel):
 </hfoption>
 <hfoption id="ResnetModelForImageClassification">
 
-The `forward` method needs to be rewrittten to calculate the loss for each logit if labels are available. Otherwise, the ResNet model class is the same.
+The `forward` method needs to be rewritten to calculate the loss for each logit if labels are available. Otherwise, the ResNet model class is the same.
 
 > [!TIP]
 > Add `config_class` to the model class to enable [AutoClass](#autoclass-support) support.

docs/source/en/gpu_selection.md

@@ -9,7 +9,7 @@ Unless required by applicable law or agreed to in writing, software distributed
 an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the
 specific language governing permissions and limitations under the License.
 
-⚠️ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be
+⚠️ Note that this file is in Markdown but contains specific syntax for our doc-builder (similar to MDX) that may not be
 rendered properly in your Markdown viewer.
 
 -->
@@ -56,7 +56,7 @@ deepspeed --num_gpus 2 trainer-program.py ...
 
 ### Order of GPUs
 
-To select specific GPUs to use and their order, configure the the `CUDA_VISIBLE_DEVICES` environment variable. It is easiest to set the environment variable in `~/bashrc` or another startup config file. `CUDA_VISIBLE_DEVICES` is used to map which GPUs are used. For example, if there are 4 GPUs (0, 1, 2, 3) and you only want to run GPUs 0 and 2:
+To select specific GPUs to use and their order, configure the `CUDA_VISIBLE_DEVICES` environment variable. It is easiest to set the environment variable in `~/bashrc` or another startup config file. `CUDA_VISIBLE_DEVICES` is used to map which GPUs are used. For example, if there are 4 GPUs (0, 1, 2, 3) and you only want to run GPUs 0 and 2:
 
 ```bash
 CUDA_VISIBLE_DEVICES=0,2 torchrun trainer-program.py ...

docs/source/es/quicktour.md

@@ -220,7 +220,7 @@ Pasa tu texto al tokenizador:
 El tokenizador devolverá un diccionario conteniendo:
 
 * [input_ids](./glossary#input-ids): representaciones numéricas de los tokens.
-* [atttention_mask](.glossary#attention-mask): indica cuáles tokens deben ser atendidos.
+* [attention_mask](.glossary#attention-mask): indica cuáles tokens deben ser atendidos.
 
 Como con el [`pipeline`], el tokenizador aceptará una lista de inputs. Además, el tokenizador también puede rellenar (pad, en inglés) y truncar el texto para devolver un lote (batch, en inglés) de longitud uniforme:
 

docs/source/it/perf_infer_cpu.md

@@ -23,7 +23,7 @@ Abbiamo integrato di recente `BetterTransformer` per fare inferenza più rapidam
 
 ## PyTorch JIT-mode (TorchScript)
 
-TorchScript è un modo di creare modelli serializzabili e ottimizzabili da codice PyTorch. Ogni programmma TorchScript può esere salvato da un processo Python  e caricato in un processo dove non ci sono dipendenze Python.
+TorchScript è un modo di creare modelli serializzabili e ottimizzabili da codice PyTorch. Ogni programma TorchScript può esere salvato da un processo Python  e caricato in un processo dove non ci sono dipendenze Python.
 Comparandolo con l'eager mode di default, jit mode in PyTorch normalmente fornisce prestazioni migliori per l'inferenza del modello da parte di metodologie di ottimizzazione come la operator fusion.
 
 Per una prima introduzione a TorchScript, vedi la Introduction to [PyTorch TorchScript tutorial](https://pytorch.org/tutorials/beginner/Intro_to_TorchScript_tutorial.html#tracing-modules).

docs/source/pt/quicktour.md

@@ -222,7 +222,7 @@ Passe o texto para o tokenizer:
 O tokenizer retornará um dicionário contendo:
 
 * [input_ids](./glossary#input-ids): representações numéricas de seus tokens.
-* [atttention_mask](.glossary#attention-mask): indica quais tokens devem ser atendidos.
+* [attention_mask](.glossary#attention-mask): indica quais tokens devem ser atendidos.
 
 Assim como o [`pipeline`], o tokenizer aceitará uma lista de entradas. Além disso, o tokenizer também pode preencher e truncar o texto para retornar um lote com comprimento uniforme:
 

src/transformers/commands/add_new_model_like.py

@@ -918,7 +918,7 @@ def add_model_to_main_init(
         new_model_patterns (`ModelPatterns`): The patterns for the new model.
         frameworks (`List[str]`, *optional*):
             If specified, only the models implemented in those frameworks will be added.
-        with_processsing (`bool`, *optional*, defaults to `True`):
+        with_processing (`bool`, *optional*, defaults to `True`):
             Whether the tokenizer/feature extractor/processor of the model should also be added to the init or not.
     """
     with open(TRANSFORMERS_PATH / "__init__.py", "r", encoding="utf-8") as f:

src/transformers/image_utils.py

@@ -94,7 +94,7 @@
     list["np.ndarray"],
     list["torch.Tensor"],
     list[list["PIL.Image.Image"]],
-    list[list["np.ndarrray"]],
+    list[list["np.ndarray"]],
     list[list["torch.Tensor"]],
 ]  # noqa
 

src/transformers/models/align/processing_align.py

@@ -83,7 +83,7 @@ def __call__(
         arguments to BertTokenizerFast's [`~BertTokenizerFast.__call__`] if `text` is not `None` to encode
         the text. To prepare the image(s), this method forwards the `images` arguments to
         EfficientNetImageProcessor's [`~EfficientNetImageProcessor.__call__`] if `images` is not `None`. Please refer
-        to the doctsring of the above two methods for more information.
+        to the docstring of the above two methods for more information.
 
         Args:
             images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):

src/transformers/models/altclip/processing_altclip.py

@@ -68,7 +68,7 @@ def __call__(
         Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
         and `kwargs` arguments to XLMRobertaTokenizerFast's [`~XLMRobertaTokenizerFast.__call__`] if `text` is not
         `None` to encode the text. To prepare the image(s), this method forwards the `images` and `kwrags` arguments to
-        CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the doctsring
+        CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the docstring
         of the above two methods for more information.
 
         Args:

src/transformers/models/auto/modeling_flax_auto.py

@@ -123,7 +123,7 @@
 
 FLAX_MODEL_FOR_IMAGE_CLASSIFICATION_MAPPING_NAMES = OrderedDict(
     [
-        # Model for Image-classsification
+        # Model for Image-classification
         ("beit", "FlaxBeitForImageClassification"),
         ("dinov2", "FlaxDinov2ForImageClassification"),
         ("regnet", "FlaxRegNetForImageClassification"),

src/transformers/models/bamba/configuration_bamba.py

@@ -39,7 +39,7 @@ class BambaConfig(PretrainedConfig):
             `inputs_ids` passed when calling [`BambaModel`]
         tie_word_embeddings (`bool`, *optional*, defaults to `False`):
             Whether the model's input and output word embeddings should be tied. Note that this is only relevant if the
-            model has a output word embedding layer.
+            model has an output word embedding layer.
         hidden_size (`int`, *optional*, defaults to 4096):
             Dimension of the hidden representations.
         intermediate_size (`int`, *optional*, defaults to 14336):
@@ -85,7 +85,7 @@ class BambaConfig(PretrainedConfig):
         mamba_n_heads (`int`, *optional*, defaults to 128):
             The number of mamba heads used in the v2 implementation.
         mamba_d_head (`int`, *optional*, defaults to `"auto"`):
-            Head embeddding dimension size
+            Head embedding dimension size
         mamba_n_groups (`int`, *optional*, defaults to 1):
             The number of the mamba groups used in the v2 implementation.
         mamba_d_state (`int`, *optional*, defaults to 256):

src/transformers/models/bark/convert_suno_to_hf.py

@@ -190,12 +190,12 @@ def load_model(pytorch_dump_folder_path, use_small=False, model_type="text"):
         output_new_model = output_new_model_total.logits[:, [-1], :]
 
     else:
-        prediction_codeboook_channel = 3
+        prediction_codebook_channel = 3
         n_codes_total = 8
         vec = torch.randint(256, (batch_size, sequence_length, n_codes_total), dtype=torch.int)
 
-        output_new_model_total = model(prediction_codeboook_channel, vec)
-        output_old_model = bark_model(prediction_codeboook_channel, vec)
+        output_new_model_total = model(prediction_codebook_channel, vec)
+        output_old_model = bark_model(prediction_codebook_channel, vec)
 
         output_new_model = output_new_model_total.logits
 

src/transformers/models/chameleon/processing_chameleon.py

@@ -87,7 +87,7 @@ def __call__(
         Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
         and `kwargs` arguments to LlamaTokenizerFast's [`~LlamaTokenizerFast.__call__`] if `text` is not `None` to encode
         the text. To prepare the image(s), this method forwards the `images` and `kwrags` arguments to
-        CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the doctsring
+        CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the docstring
         of the above two methods for more information.
 
         Args:

src/transformers/models/chinese_clip/processing_chinese_clip.py

@@ -78,7 +78,7 @@ def __call__(
         Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
         and `kwargs` arguments to BertTokenizerFast's [`~BertTokenizerFast.__call__`] if `text` is not `None` to encode
         the text. To prepare the image(s), this method forwards the `images` and `kwrags` arguments to
-        CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the doctsring
+        CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the docstring
         of the above two methods for more information.
 
         Args:

src/transformers/models/clap/processing_clap.py

@@ -46,7 +46,7 @@ def __call__(self, text=None, audios=None, return_tensors=None, **kwargs):
         and `kwargs` arguments to RobertaTokenizerFast's [`~RobertaTokenizerFast.__call__`] if `text` is not `None` to
         encode the text. To prepare the audio(s), this method forwards the `audios` and `kwrags` arguments to
         ClapFeatureExtractor's [`~ClapFeatureExtractor.__call__`] if `audios` is not `None`. Please refer to the
-        doctsring of the above two methods for more information.
+        docstring of the above two methods for more information.
 
         Args:
             text (`str`, `List[str]`, `List[List[str]]`):

src/transformers/models/clip/processing_clip.py

@@ -63,7 +63,7 @@ def __call__(self, text=None, images=None, return_tensors=None, **kwargs):
         Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
         and `kwargs` arguments to CLIPTokenizerFast's [`~CLIPTokenizerFast.__call__`] if `text` is not `None` to encode
         the text. To prepare the image(s), this method forwards the `images` and `kwrags` arguments to
-        CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the doctsring
+        CLIPImageProcessor's [`~CLIPImageProcessor.__call__`] if `images` is not `None`. Please refer to the docstring
         of the above two methods for more information.
 
         Args:

src/transformers/models/clipseg/processing_clipseg.py

@@ -63,7 +63,7 @@ def __call__(self, text=None, images=None, visual_prompt=None, return_tensors=No
         Main method to prepare for the model one or several sequences(s) and image(s). This method forwards the `text`
         and `kwargs` arguments to CLIPTokenizerFast's [`~CLIPTokenizerFast.__call__`] if `text` is not `None` to encode
         the text. To prepare the image(s), this method forwards the `images` and `kwrags` arguments to
-        ViTImageProcessor's [`~ViTImageProcessor.__call__`] if `images` is not `None`. Please refer to the doctsring of
+        ViTImageProcessor's [`~ViTImageProcessor.__call__`] if `images` is not `None`. Please refer to the docstring of
         the above two methods for more information.
 
         Args:

src/transformers/models/clvp/processing_clvp.py

@@ -48,7 +48,7 @@ def __init__(self, feature_extractor, tokenizer):
     def __call__(self, *args, **kwargs):
         """
         Forwards the `audio` and `sampling_rate` arguments to [`~ClvpFeatureExtractor.__call__`] and the `text`
-        argument to [`~ClvpTokenizer.__call__`]. Please refer to the doctsring of the above two methods for more
+        argument to [`~ClvpTokenizer.__call__`]. Please refer to the docstring of the above two methods for more
         information.
         """
 

src/transformers/models/colpali/modular_colpali.py

@@ -100,11 +100,11 @@ def __call__(
         wrapper around the PaliGemmaProcessor's [`~PaliGemmaProcessor.__call__`] method adapted for the ColPali model. It cannot process
         both text and images at the same time.
 
-        When preparing the the text(s), this method forwards the `text` and `kwargs` arguments to LlamaTokenizerFast's
+        When preparing the text(s), this method forwards the `text` and `kwargs` arguments to LlamaTokenizerFast's
         [`~LlamaTokenizerFast.__call__`].
-        When preparing the the image(s), this method forwards the `images` and `kwargs` arguments to SiglipImageProcessor's
+        When preparing the image(s), this method forwards the `images` and `kwargs` arguments to SiglipImageProcessor's
         [`~SiglipImageProcessor.__call__`].
-        Please refer to the doctsring of the above two methods for more information.
+        Please refer to the docstring of the above two methods for more information.
 
         Args:
             images (`PIL.Image.Image`, `np.ndarray`, `torch.Tensor`, `List[PIL.Image.Image]`, `List[np.ndarray]`, `List[torch.Tensor]`):