[docs] fixed links with 404 (#27327)

* fixed links with 404

* make style

docs/source/ja/perf_train_gpu_one.md

@@ -323,7 +323,7 @@ training_args = TrainingArguments(torch_compile=True, **default_args)
 
 **推論専用バックエンド**：
 * `dynamo.optimize("ofi")` - Torchscriptの`optimize_for_inference`を使用します。 [詳細はこちら](https://pytorch.org/docs/stable/generated/torch.jit.optimize_for_inference.html)
-* `dynamo.optimize("fx2trt")` - Nvidia TensorRTを使用した推論の最適化にNvidia TensorRTを使用します。 [詳細はこちら](https://github.com/pytorch/TensorRT/blob/master/docsrc/tutorials/getting_started_with_fx_path.rst)
+* `dynamo.optimize("fx2trt")` - Nvidia TensorRTを使用した推論の最適化にNvidia TensorRTを使用します。 [詳細はこちら](https://pytorch.org/TensorRT/tutorials/getting_started_with_fx_path.html)
 * `dynamo.optimize("onnxrt")` - CPU/GPUでの推論にONNX Runtimeを使用します。 [詳細はこちら](https://onnxruntime.ai/)
 * `dynamo.optimize("ipex")` - CPUでの推論にIPEXを使用します。 [詳細はこちら](https://github.com/intel/intel-extension-for-pytorch)
 
@@ -411,7 +411,7 @@ PyTorchの[pipとcondaビルド](https://pytorch.org/get-started/locally/#start-
 - [Switch Transformers: シンプルで効率的なスパース性を備えたトリリオンパラメータモデルへのスケーリング](https://arxiv.org/abs/2101.03961)
 - [GLaM: Generalist Language Model (GLaM)](https://ai.googleblog.com/2021/12/more-efficient-in-context-learning-with.html)
 
-PytorchにはDeepSpeedが構築したものもあります: [DeepSpeed-MoE: Advancing Mixture-of-Experts Inference and Training to Power Next-Generation AI Scale](https://arxiv.org/abs/2201.05596)、[Mixture of Experts](https://www.deepspeed.ai/tutorials/mixture-of-experts/) - ブログ記事: [1](https://www.microsoft.com/en-us/research/blog/deepspeed-powers-8x-larger-moe-model-training-with-high-performance/)、[2](https://www.microsoft.com/en-us/research/publication/scalable-and-efficient-moe-training-for-multitask-multilingual-models/)、大規模なTransformerベースの自然言語生成モデルの具体的な展開については、[ブログ記事](https://www.deepspeed.ai/news/2021/12/09/deepspeed-moe-nlg.html)、[Megatron-Deepspeedブランチ](Thttps://github.com/microsoft/Megatron-DeepSpeed/tree/moe-training)を参照してください。
+PytorchにはDeepSpeedが構築したものもあります: [DeepSpeed-MoE: Advancing Mixture-of-Experts Inference and Training to Power Next-Generation AI Scale](https://arxiv.org/abs/2201.05596)、[Mixture of Experts](https://www.deepspeed.ai/tutorials/mixture-of-experts/) - ブログ記事: [1](https://www.microsoft.com/en-us/research/blog/deepspeed-powers-8x-larger-moe-model-training-with-high-performance/)、[2](https://www.microsoft.com/en-us/research/publication/scalable-and-efficient-moe-training-for-multitask-multilingual-models/)、大規模なTransformerベースの自然言語生成モデルの具体的な展開については、[ブログ記事](https://www.deepspeed.ai/2021/12/09/deepspeed-moe-nlg.html)、[Megatron-Deepspeedブランチ](https://github.com/microsoft/Megatron-DeepSpeed/tree/moe-training)を参照してください。
 
 
 ## PyTorchネイティブアテンションとFlash Attentionの使用

docs/source/ja/philosophy.md

@@ -56,7 +56,7 @@ rendered properly in your Markdown viewer.
 
 このライブラリは、各モデルについて次の3つのタイプのクラスを中心に構築されています：
 
-- **モデルクラス**は、ライブラリで提供される事前トレーニング済みの重みと互換性のあるPyTorchモデル（[torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module)）、Kerasモデル（[tf.keras.Model](https://www.tensorflow.org/api_docs/python/tf/keras/Model)）またはJAX/Flaxモデル（[flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen.html)）を使用できます。
+- **モデルクラス**は、ライブラリで提供される事前トレーニング済みの重みと互換性のあるPyTorchモデル（[torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module)）、Kerasモデル（[tf.keras.Model](https://www.tensorflow.org/api_docs/python/tf/keras/Model)）またはJAX/Flaxモデル（[flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html)）を使用できます。
 - **構成クラス**は、モデルを構築するために必要なハイパーパラメータを格納します（層の数や隠れ層のサイズなど）。これらを自分でインスタンス化する必要はありません。特に、変更を加えずに事前トレーニング済みモデルを使用している場合、モデルを作成すると自動的に構成がインスタンス化されるようになります（これはモデルの一部です）。
 - **前処理クラス**は、生データをモデルが受け入れる形式に変換します。[トークナイザ](main_classes/tokenizer)は各モデルの語彙を保存し、文字列をトークン埋め込みのインデックスのリストにエンコードおよびデコードするためのメソッドを提供します。[イメージプロセッサ](main_classes/image_processor)はビジョン入力を前処理し、[特徴抽出器](main_classes/feature_extractor)はオーディオ入力を前処理し、[プロセッサ](main_classes/processors)はマルチモーダル入力を処理します。
 

docs/source/ko/create_a_model.md

@@ -110,7 +110,7 @@ configuration 파일을 딕셔너리로 저장하거나 사용자 정의 configu
 
 ## 모델[[model]]
 
-다음 단계는 [모델(model)](main_classes/models)을 만드는 것입니다. 느슨하게 아키텍처라고도 불리는 모델은 각 계층이 수행하는 동작과 발생하는 작업을 정의합니다. configuration의 `num_hidden_layers`와 같은 속성은 아키텍처를 정의하는 데 사용됩니다. 모든 모델은 기본 클래스 [`PreTrainedModel`]과 입력 임베딩 크기 조정 및 셀프 어텐션 헤드 가지 치기와 같은 몇 가지 일반적인 메소드를 공유합니다. 또한 모든 모델은 [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html), [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) 또는 [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/flax.linen.html#module)의 서브클래스이기도 합니다. 즉, 모델은 각 프레임워크의 사용법과 호환됩니다.
+다음 단계는 [모델(model)](main_classes/models)을 만드는 것입니다. 느슨하게 아키텍처라고도 불리는 모델은 각 계층이 수행하는 동작과 발생하는 작업을 정의합니다. configuration의 `num_hidden_layers`와 같은 속성은 아키텍처를 정의하는 데 사용됩니다. 모든 모델은 기본 클래스 [`PreTrainedModel`]과 입력 임베딩 크기 조정 및 셀프 어텐션 헤드 가지 치기와 같은 몇 가지 일반적인 메소드를 공유합니다. 또한 모든 모델은 [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html), [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) 또는 [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html)의 서브클래스이기도 합니다. 즉, 모델은 각 프레임워크의 사용법과 호환됩니다.
 
 <frameworkcontent>
 <pt>

docs/source/ko/perf_train_gpu_many.md

@@ -240,7 +240,7 @@ ZeRO가 모델 가중치를 분할하는 방식을 자세히 살펴보면, 텐
 
 구현:
 
-- [DeepSpeed](https://www.deepspeed.ai/features/#the-zero-redundancy-optimizer)는 1단계 + 2단계 + 3단계의 ZeRO-DP를 제공합니다.
+- [DeepSpeed](https://www.deepspeed.ai/tutorials/zero/)는 1단계 + 2단계 + 3단계의 ZeRO-DP를 제공합니다.
 - [Fairscale](https://github.com/facebookresearch/fairscale/#optimizer-state-sharding-zero)은 1단계 + 2단계 + 3단계의 ZeRO-DP를 제공합니다.
 - [`transformers` 통합](main_classes/trainer#trainer-integrations)
 
@@ -362,7 +362,7 @@ Megatron 논문의 표기법을 따라 행렬의 점곱 부분을 `Y = GeLU(XA)`
 SageMaker는 더 효율적인 처리를 위해 TP와 DP를 결합합니다.
 
 대체 이름:
-- DeepSpeed는 이를 [텐서 슬라이싱](https://www.deepspeed.ai/features/#model-parallelism)이라고 부릅니다.
+- DeepSpeed는 이를 [텐서 슬라이싱](https://www.deepspeed.ai/training/#model-parallelism)이라고 부릅니다.
 
 구현:
 - [Megatron-LM](https://github.com/NVIDIA/Megatron-LM)은 내부 구현을 가지고 있으므로 모델에 매우 특화되어 있습니다.

docs/source/ko/philosophy.md

@@ -54,7 +54,7 @@ rendered properly in your Markdown viewer.
 
 이 라이브러리는 각 모델에 대해 세 가지 유형의 클래스를 기반으로 구축되었습니다:
 
-- **모델 클래스**는 라이브러리에서 제공하는 사전 훈련된 가중치와 함께 작동하는 PyTorch 모델([torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module)), Keras 모델([tf.keras.Model](https://www.tensorflow.org/api_docs/python/tf/keras/Model)), JAX/Flax 모델([flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen.html))일 수 있습니다.
+- **모델 클래스**는 라이브러리에서 제공하는 사전 훈련된 가중치와 함께 작동하는 PyTorch 모델([torch.nn.Module](https://pytorch.org/docs/stable/nn.html#torch.nn.Module)), Keras 모델([tf.keras.Model](https://www.tensorflow.org/api_docs/python/tf/keras/Model)), JAX/Flax 모델([flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html))일 수 있습니다.
 - **구성 클래스**는 모델을 구축하는 데 필요한 하이퍼파라미터(예: 레이어 수 및 은닉 크기)를 저장합니다. 구성 클래스를 직접 인스턴스화할 필요는 없습니다. 특히, 수정 없이 고 사전 학습된 모델을 사용하는 경우 모델을 생성하면 모델의 일부인 구성을 자동으로 인스턴스화됩니다.
 - **전처리 클래스**는 원시 데이터를 모델이 수용하는 형식으로 변환합니다. [Tokenizer](main_classes/tokenizer)는 각 모델의 어휘를 저장하고, 문자열을 토큰 임베딩 인덱스 리스트로 인코딩하고 디코딩하기 위한 메소드를 제공합니다. [Image processors](main_classes/image_processor)는 비전 입력을 전처리하고, [feature extractors](main_classes/feature_extractor)는 오디오 입력을 전처리하며, [processor](main_classes/processors)는 멀티모달 입력을 처리합니다.
 

docs/source/ko/sagemaker.md

@@ -26,4 +26,3 @@ rendered properly in your Markdown viewer.
 
 - [Train Hugging Face models on Amazon SageMaker with the SageMaker Python SDK](https://huggingface.co/docs/sagemaker/train)
 - [Deploy Hugging Face models to Amazon SageMaker with the SageMaker Python SDK](https://huggingface.co/docs/sagemaker/inference)
-- [Frequently Asked Questions](https://huggingface.co/docs/sagemaker/faq)

docs/source/pt/create_a_model.md

@@ -109,7 +109,7 @@ Você pode também salvar seu arquivo de configurações como um dicionário ou
 
 ## Modelo
 
-O próximo passo é criar um [model](main_classes/models). O modelo - também vagamente referido como arquitetura - define o que cada camada está fazendo e quais operações estão acontecendo. Atributos como `num_hidden_layers` das configurações são utilizados para definir a arquitetura. Todo modelo compartilha a classe base [`PreTrainedModel`] e alguns métodos em comum como redimensionar o tamanho dos embeddings de entrada e podar as 'self-attention heads'. Além disso, todos os modelos também são subclasses de [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html), [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) ou [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/flax.linen.html#module). Isso significa que os modelos são compatíveis com cada respectivo uso de framework.
+O próximo passo é criar um [model](main_classes/models). O modelo - também vagamente referido como arquitetura - define o que cada camada está fazendo e quais operações estão acontecendo. Atributos como `num_hidden_layers` das configurações são utilizados para definir a arquitetura. Todo modelo compartilha a classe base [`PreTrainedModel`] e alguns métodos em comum como redimensionar o tamanho dos embeddings de entrada e podar as 'self-attention heads'. Além disso, todos os modelos também são subclasses de [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html), [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) ou [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html). Isso significa que os modelos são compatíveis com cada respectivo uso de framework.
 
 <frameworkcontent>
 <pt>

docs/source/zh/create_a_model.md

@@ -110,7 +110,7 @@ DistilBertConfig {
 
 ## 模型
 
-接下来，创建一个[模型](main_classes/models)。模型，也可泛指架构，定义了每一层网络的行为以及进行的操作。配置中的 `num_hidden_layers` 等属性用于定义架构。每个模型都共享基类 [`PreTrainedModel`] 和一些常用方法，例如调整输入嵌入的大小和修剪自注意力头。此外，所有模型都是 [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html)、[`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) 或 [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/flax.linen.html#module) 的子类。这意味着模型与各自框架的用法兼容。
+接下来，创建一个[模型](main_classes/models)。模型，也可泛指架构，定义了每一层网络的行为以及进行的操作。配置中的 `num_hidden_layers` 等属性用于定义架构。每个模型都共享基类 [`PreTrainedModel`] 和一些常用方法，例如调整输入嵌入的大小和修剪自注意力头。此外，所有模型都是 [`torch.nn.Module`](https://pytorch.org/docs/stable/generated/torch.nn.Module.html)、[`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model) 或 [`flax.linen.Module`](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) 的子类。这意味着模型与各自框架的用法兼容。
 
 <frameworkcontent>
 <pt>

src/transformers/models/albert/modeling_flax_albert.py

@@ -86,9 +86,10 @@ ALBERT_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/beit/modeling_flax_beit.py

@@ -69,9 +69,10 @@ BEIT_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/bert/modeling_flax_bert.py

@@ -93,9 +93,10 @@ BERT_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/big_bird/modeling_flax_big_bird.py

@@ -122,9 +122,10 @@ BIG_BIRD_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/clip/modeling_flax_clip.py

@@ -43,9 +43,10 @@ CLIP_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/distilbert/modeling_flax_distilbert.py

@@ -48,9 +48,10 @@ FLAX_DISTILBERT_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/regnet/modeling_flax_regnet.py

@@ -47,9 +47,10 @@ REGNET_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/resnet/modeling_flax_resnet.py

@@ -42,9 +42,10 @@ RESNET_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/roberta/modeling_flax_roberta.py

@@ -78,9 +78,10 @@ ROBERTA_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/roberta_prelayernorm/modeling_flax_roberta_prelayernorm.py

@@ -80,9 +80,10 @@ ROBERTA_PRELAYERNORM_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/roformer/modeling_flax_roformer.py

@@ -59,9 +59,10 @@ ROFORMER_START_DOCSTRING = r"""
     This model inherits from [`FlaxPreTrainedModel`]. Check the superclass documentation for the generic methods the
     library implements for all its model (such as downloading, saving and converting weights from PyTorch models)
 
-    This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-    subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-    general usage and behavior.
+    This model is also a
+    [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it as
+    a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+    behavior.
 
     Finally, this model supports inherent JAX features such as:
 

src/transformers/models/vision_text_dual_encoder/modeling_flax_vision_text_dual_encoder.py

@@ -52,9 +52,10 @@ VISION_TEXT_DUAL_ENCODER_START_DOCSTRING = r"""
     library implements for all its model (such as downloading or saving, resizing the input embeddings, pruning heads
     etc.)
 
-     This model is also a Flax Linen [flax.linen.Module](https://flax.readthedocs.io/en/latest/flax.linen.html#module)
-     subclass. Use it as a regular Flax linen Module and refer to the Flax documentation for all matter related to
-     general usage and behavior.
+     This model is also a
+     [flax.linen.Module](https://flax.readthedocs.io/en/latest/api_reference/flax.linen/module.html) subclass. Use it
+     as a regular Flax linen Module and refer to the Flax documentation for all matter related to general usage and
+     behavior.
 
     Finally, this model supports inherent JAX features such as: