translation main-class files to chinese (#27588)

* translate work

* update

* update

* update [[autodoc]]

* Update callback.md

---------
Co-authored-by: jiaqiw <wangjiaqi50@huawei.com>

docs/source/zh/_toctree.yml

@@ -68,11 +68,46 @@
   title: 概念指南
 - sections:
   - sections:
+    - local: main_classes/agent
+      title: Agents和工具
+    - local: main_classes/callback
+      title: Callbacks
+    - local: main_classes/configuration
+      title: Configuration
+    - local: main_classes/data_collator
+      title: Data Collator
+    - local: main_classes/keras_callbacks
+      title: Keras callbacks
+    - local: main_classes/logging
+      title: Logging
     - local: main_classes/model
       title: 模型
+    - local: main_classes/text_generation
+      title: 文本生成
+    - local: main_classes/onnx
+      title: ONNX
+    - local: main_classes/optimizer_schedules
+      title: Optimization
+    - local: main_classes/output
+      title: 模型输出
+    - local: main_classes/pipelines
+      title: Pipelines
+    - local: main_classes/processors
+      title: Processors
+    - local: main_classes/quantization
+      title: Quantization
+    - local: main_classes/tokenizer
+      title: Tokenizer
     - local: main_classes/trainer
       title: Trainer
     - local: main_classes/deepspeed
       title: DeepSpeed集成
+    - local: main_classes/feature_extractor
+      title: Feature Extractor
+    - local: main_classes/image_processor
+      title: Image Processor
     title: 主要类
   title: 应用程序接口 (API) 
+
+
+

docs/source/zh/main_classes/agent.md

+<!--Copyright 2023 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Agents和工具
+
+<Tip warning={true}>
+
+Transformers Agents是一个实验性的API，它随时可能发生变化。由于API或底层模型容易发生变化，因此由agents返回的结果可能会有所不同。
+
+
+</Tip>
+
+要了解更多关于agents和工具的信息，请确保阅读[介绍指南](../transformers_agents)。此页面包含底层类的API文档。
+
+
+## Agents
+
+我们提供三种类型的agents：[`HfAgent`]使用开源模型的推理端点，[`LocalAgent`]使用您在本地选择的模型，[`OpenAiAgent`]使用OpenAI封闭模型。
+
+
+### HfAgent
+
+[[autodoc]] HfAgent
+
+### LocalAgent
+
+[[autodoc]] LocalAgent
+
+### OpenAiAgent
+
+[[autodoc]] OpenAiAgent
+
+### AzureOpenAiAgent
+
+[[autodoc]] AzureOpenAiAgent
+
+### Agent
+
+[[autodoc]] Agent 
+    - chat 
+    - run 
+    - prepare_for_new_chat
+
+## 工具
+
+### load_tool
+
+[[autodoc]] load_tool
+
+### Tool
+
+[[autodoc]] Tool
+
+### PipelineTool
+
+[[autodoc]] PipelineTool
+
+### RemoteTool
+
+[[autodoc]] RemoteTool
+
+### launch_gradio_demo
+
+[[autodoc]] launch_gradio_demo
+
+## Agent类型
+
+Agents可以处理工具之间任何类型的对象；工具是多模态的，可以接受和返回文本、图像、音频、视频等类型。为了增加工具之间的兼容性，以及正确地在ipython（jupyter、colab、ipython notebooks等）中呈现这些返回值，我们实现了这些类型的包装类。
+
+被包装的对象应该继续按照最初的行为方式运作；文本对象应该仍然像字符串一样运作，图像对象应该仍然像`PIL.Image`一样运作。
+
+这些类型有三个特定目的：
+
+- 对类型调用 `to_raw` 应该返回底层对象
+- 对类型调用 `to_string` 应该将对象作为字符串返回：在`AgentText`的情况下可能是字符串，但在其他情况下可能是对象序列化版本的路径
+- 在ipython内核中显示它应该正确显示对象
+
+### AgentText
+
+[[autodoc]] transformers.tools.agent_types.AgentText
+
+### AgentImage
+
+[[autodoc]] transformers.tools.agent_types.AgentImage
+
+### AgentAudio
+
+[[autodoc]] transformers.tools.agent_types.AgentAudio

docs/source/zh/main_classes/callback.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Callbacks
+
+
+Callbacks可以用来自定义PyTorch [Trainer]中训练循环行为的对象（此功能尚未在TensorFlow中实现），该对象可以检查训练循环状态（用于进度报告、在TensorBoard或其他ML平台上记录日志等），并做出决策（例如提前停止）。
+
+Callbacks是“只读”的代码片段，除了它们返回的[TrainerControl]对象外，它们不能更改训练循环中的任何内容。对于需要更改训练循环的自定义，您应该继承[Trainer]并重载您需要的方法（有关示例，请参见[trainer](trainer)）。
+
+默认情况下，`TrainingArguments.report_to` 设置为"all"，然后[Trainer]将使用以下callbacks。
+
+
+- [`DefaultFlowCallback`]，它处理默认的日志记录、保存和评估行为
+- [`PrinterCallback`] 或 [`ProgressCallback`]，用于显示进度和打印日志（如果通过[`TrainingArguments`]停用tqdm，则使用第一个函数；否则使用第二个）。
+- [`~integrations.TensorBoardCallback`]，如果TensorBoard可访问（通过PyTorch版本 >= 1.4 或者 tensorboardX）。
+- [`~integrations.WandbCallback`]，如果安装了[wandb](https://www.wandb.com/)。
+- [`~integrations.CometCallback`]，如果安装了[comet_ml](https://www.comet.ml/site/)。
+- [`~integrations.MLflowCallback`]，如果安装了[mlflow](https://www.mlflow.org/)。
+- [`~integrations.NeptuneCallback`]，如果安装了[neptune](https://neptune.ai/)。
+- [`~integrations.AzureMLCallback`]，如果安装了[azureml-sdk](https://pypi.org/project/azureml-sdk/)。
+- [`~integrations.CodeCarbonCallback`]，如果安装了[codecarbon](https://pypi.org/project/codecarbon/)。
+- [`~integrations.ClearMLCallback`]，如果安装了[clearml](https://github.com/allegroai/clearml)。
+- [`~integrations.DagsHubCallback`]，如果安装了[dagshub](https://dagshub.com/)。
+- [`~integrations.FlyteCallback`]，如果安装了[flyte](https://flyte.org/)。
+- [`~integrations.DVCLiveCallback`]，如果安装了[dvclive](https://dvc.org/doc/dvclive)。
+
+如果安装了一个软件包，但您不希望使用相关的集成，您可以将 `TrainingArguments.report_to` 更改为仅包含您想要使用的集成的列表（例如 `["azure_ml", "wandb"]`）。
+
+实现callbacks的主要类是[`TrainerCallback`]。它获取用于实例化[`Trainer`]的[`TrainingArguments`]，可以通过[`TrainerState`]访问该Trainer的内部状态，并可以通过[`TrainerControl`]对训练循环执行一些操作。
+
+
+## 可用的Callbacks
+
+这里是库里可用[`TrainerCallback`]的列表：
+
+[[autodoc]] integrations.CometCallback 
+    - setup
+
+[[autodoc]] DefaultFlowCallback
+
+[[autodoc]] PrinterCallback
+
+[[autodoc]] ProgressCallback
+
+[[autodoc]] EarlyStoppingCallback
+
+[[autodoc]] integrations.TensorBoardCallback
+
+[[autodoc]] integrations.WandbCallback 
+    - setup
+
+[[autodoc]] integrations.MLflowCallback 
+    - setup
+
+[[autodoc]] integrations.AzureMLCallback
+
+[[autodoc]] integrations.CodeCarbonCallback
+
+[[autodoc]] integrations.NeptuneCallback
+
+[[autodoc]] integrations.ClearMLCallback
+
+[[autodoc]] integrations.DagsHubCallback
+
+[[autodoc]] integrations.FlyteCallback
+
+[[autodoc]] integrations.DVCLiveCallback
+    - setup
+
+## TrainerCallback
+
+[[autodoc]] TrainerCallback
+
+以下是如何使用PyTorch注册自定义callback的示例：
+
+[`Trainer`]:
+
+```python
+class MyCallback(TrainerCallback):
+    "A callback that prints a message at the beginning of training"
+
+    def on_train_begin(self, args, state, control, **kwargs):
+        print("Starting training")
+
+
+trainer = Trainer(
+    model,
+    args,
+    train_dataset=train_dataset,
+    eval_dataset=eval_dataset,
+    callbacks=[MyCallback],  # We can either pass the callback class this way or an instance of it (MyCallback())
+)
+```
+
+注册callback的另一种方式是调用 `trainer.add_callback()`，如下所示：
+
+
+```python
+trainer = Trainer(...)
+trainer.add_callback(MyCallback)
+# Alternatively, we can pass an instance of the callback class
+trainer.add_callback(MyCallback())
+```
+
+## TrainerState
+
+[[autodoc]] TrainerState
+
+## TrainerControl
+
+[[autodoc]] TrainerControl

docs/source/zh/main_classes/configuration.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Configuration
+
+基类[`PretrainedConfig`]实现了从本地文件或目录加载/保存配置的常见方法，或下载库提供的预训练模型配置（从HuggingFace的AWS S3库中下载）。
+
+每个派生的配置类都实现了特定于模型的属性。所有配置类中共同存在的属性有：`hidden_size`、`num_attention_heads` 和 `num_hidden_layers`。文本模型进一步添加了 `vocab_size`。
+
+
+## PretrainedConfig
+
+[[autodoc]] PretrainedConfig
+    - push_to_hub
+    - all

docs/source/zh/main_classes/data_collator.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Data Collator
+
+Data collators是一个对象，通过使用数据集元素列表作为输入来形成一个批次。这些元素与 `train_dataset` 或 `eval_dataset` 的元素类型相同。
+
+为了能够构建批次，Data collators可能会应用一些预处理（比如填充）。其中一些（比如[`DataCollatorForLanguageModeling`]）还会在形成的批次上应用一些随机数据增强（比如随机掩码）。
+
+在[示例脚本](../examples)或[示例notebooks](../notebooks)中可以找到使用的示例。
+
+
+## Default data collator
+
+[[autodoc]] data.data_collator.default_data_collator
+
+## DefaultDataCollator
+
+[[autodoc]] data.data_collator.DefaultDataCollator
+
+## DataCollatorWithPadding
+
+[[autodoc]] data.data_collator.DataCollatorWithPadding
+
+## DataCollatorForTokenClassification
+
+[[autodoc]] data.data_collator.DataCollatorForTokenClassification
+
+## DataCollatorForSeq2Seq
+
+[[autodoc]] data.data_collator.DataCollatorForSeq2Seq
+
+## DataCollatorForLanguageModeling
+
+[[autodoc]] data.data_collator.DataCollatorForLanguageModeling
+    - numpy_mask_tokens
+    - tf_mask_tokens
+    - torch_mask_tokens
+
+## DataCollatorForWholeWordMask
+
+[[autodoc]] data.data_collator.DataCollatorForWholeWordMask
+    - numpy_mask_tokens
+    - tf_mask_tokens
+    - torch_mask_tokens
+
+## DataCollatorForPermutationLanguageModeling
+
+[[autodoc]] data.data_collator.DataCollatorForPermutationLanguageModeling
+    - numpy_mask_tokens
+    - tf_mask_tokens
+    - torch_mask_tokens

docs/source/zh/main_classes/feature_extractor.md

+<!--Copyright 2021 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Feature Extractor
+
+Feature Extractor负责为音频或视觉模型准备输入特征。这包括从序列中提取特征，例如，对音频文件进行预处理以生成Log-Mel频谱特征，以及从图像中提取特征，例如，裁剪图像文件，同时还包括填充、归一化和转换为NumPy、PyTorch和TensorFlow张量。
+
+
+## FeatureExtractionMixin
+
+[[autodoc]] feature_extraction_utils.FeatureExtractionMixin
+    - from_pretrained
+    - save_pretrained
+
+## SequenceFeatureExtractor
+
+[[autodoc]] SequenceFeatureExtractor
+    - pad
+
+## BatchFeature
+
+[[autodoc]] BatchFeature
+
+## ImageFeatureExtractionMixin
+
+[[autodoc]] image_utils.ImageFeatureExtractionMixin

docs/source/zh/main_classes/image_processor.md

+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Image Processor
+
+Image processor负责为视觉模型准备输入特征并后期处理处理它们的输出。这包括诸如调整大小、归一化和转换为PyTorch、TensorFlow、Flax和NumPy张量等转换。它还可能包括特定于模型的后期处理，例如将logits转换为分割掩码。
+
+
+## ImageProcessingMixin
+
+[[autodoc]] image_processing_utils.ImageProcessingMixin
+    - from_pretrained
+    - save_pretrained
+
+## BatchFeature
+
+[[autodoc]] BatchFeature
+
+## BaseImageProcessor
+
+[[autodoc]] image_processing_utils.BaseImageProcessor

docs/source/zh/main_classes/keras_callbacks.md

+<!--Copyright 2021 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Keras callbacks
+
+在Keras中训练Transformers模型时，有一些库特定的callbacks函数可用于自动执行常见任务：
+
+## KerasMetricCallback
+
+[[autodoc]] KerasMetricCallback
+
+## PushToHubCallback
+
+[[autodoc]] PushToHubCallback

docs/source/zh/main_classes/logging.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Logging
+
+🤗 Transformers拥有一个集中式的日志系统，因此您可以轻松设置库输出的日志详细程度。
+
+当前库的默认日志详细程度为`WARNING`。
+
+要更改日志详细程度，只需使用其中一个直接的setter。例如，以下是如何将日志详细程度更改为INFO级别的方法：
+
+```python
+import transformers
+
+transformers.logging.set_verbosity_info()
+```
+
+您还可以使用环境变量`TRANSFORMERS_VERBOSITY`来覆盖默认的日志详细程度。您可以将其设置为以下级别之一：`debug`、`info`、`warning`、`error`、`critical`。例如：
+
+```bash
+TRANSFORMERS_VERBOSITY=error ./myprogram.py
+```
+
+此外，通过将环境变量`TRANSFORMERS_NO_ADVISORY_WARNINGS`设置为`true`（如*1*），可以禁用一些`warnings`。这将禁用[`logger.warning_advice`]记录的任何警告。例如：
+
+```bash
+TRANSFORMERS_NO_ADVISORY_WARNINGS=1 ./myprogram.py
+```
+
+以下是如何在您自己的模块或脚本中使用与库相同的logger的示例：
+
+```python
+from transformers.utils import logging
+
+logging.set_verbosity_info()
+logger = logging.get_logger("transformers")
+logger.info("INFO")
+logger.warning("WARN")
+```
+
+
+此日志模块的所有方法都在下面进行了记录，主要的方法包括 [`logging.get_verbosity`] 用于获取logger当前输出日志详细程度的级别和 [`logging.set_verbosity`] 用于将详细程度设置为您选择的级别。按照顺序（从最不详细到最详细），这些级别（及其相应的整数值）为：
+
+- `transformers.logging.CRITICAL` 或 `transformers.logging.FATAL`（整数值，50）：仅报告最关键的errors。
+- `transformers.logging.ERROR`（整数值，40）：仅报告errors。
+- `transformers.logging.WARNING` 或 `transformers.logging.WARN`（整数值，30）：仅报告error和warnings。这是库使用的默认级别。
+- `transformers.logging.INFO`（整数值，20）：报告error、warnings和基本信息。
+- `transformers.logging.DEBUG`（整数值，10）：报告所有信息。
+
+默认情况下，将在模型下载期间显示`tqdm`进度条。[`logging.disable_progress_bar`] 和 [`logging.enable_progress_bar`] 可用于禁止或启用此行为。
+
+## `logging` vs `warnings`
+
+Python有两个经常一起使用的日志系统：如上所述的`logging`，和对特定buckets中的警告进行进一步分类的`warnings`，例如，`FutureWarning`用于输出已经被弃用的功能或路径，`DeprecationWarning`用于指示即将被弃用的内容。
+
+我们在`transformers`库中同时使用这两个系统。我们利用并调整了`logging`的`captureWarning`方法，以便通过上面的详细程度setters来管理这些警告消息。
+
+对于库的开发人员，这意味着什么呢？我们应该遵循以下启发法则：
+- 库的开发人员和依赖于`transformers`的库应优先使用`warnings`
+- `logging`应该用于在日常项目中经常使用它的用户
+
+以下是`captureWarnings`方法的参考。
+
+[[autodoc]] logging.captureWarnings
+
+## Base setters
+
+[[autodoc]] logging.set_verbosity_error
+
+[[autodoc]] logging.set_verbosity_warning
+
+[[autodoc]] logging.set_verbosity_info
+
+[[autodoc]] logging.set_verbosity_debug
+
+## Other functions
+
+[[autodoc]] logging.get_verbosity
+
+[[autodoc]] logging.set_verbosity
+
+[[autodoc]] logging.get_logger
+
+[[autodoc]] logging.enable_default_handler
+
+[[autodoc]] logging.disable_default_handler
+
+[[autodoc]] logging.enable_explicit_format
+
+[[autodoc]] logging.reset_format
+
+[[autodoc]] logging.enable_progress_bar
+
+[[autodoc]] logging.disable_progress_bar

docs/source/zh/main_classes/model.md

@@ -119,16 +119,16 @@ model = AutoModel.from_config(config)
 
 TFPreTrainedModel
 [[autodoc]] TFPreTrainedModel
-- push_to_hub
-- all
+    - push_to_hub
+    - all
 
 ## TFModelUtilsMixin
 [[autodoc]] modeling_tf_utils.TFModelUtilsMixin
 
 FlaxPreTrainedModel
 [[autodoc]] FlaxPreTrainedModel
-- push_to_hub
-- all
+    - push_to_hub
+    - all
 
 ## 推送到 Hub
 [[autodoc]] utils.PushToHubMixin

docs/source/zh/main_classes/onnx.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# 导出 🤗 Transformers 模型到 ONNX
+
+🤗 Transformers提供了一个`transformers.onnx`包，通过利用配置对象，您可以将模型checkpoints转换为ONNX图。
+
+有关更多详细信息，请参阅导出 🤗 Transformers 模型的[指南](../serialization)。
+
+## ONNX Configurations
+
+我们提供了三个抽象类，取决于您希望导出的模型架构类型：
+
+* 基于编码器的模型继承 [`~onnx.config.OnnxConfig`]
+* 基于解码器的模型继承 [`~onnx.config.OnnxConfigWithPast`]
+* 编码器-解码器模型继承 [`~onnx.config.OnnxSeq2SeqConfigWithPast`]
+
+### OnnxConfig
+
+[[autodoc]] onnx.config.OnnxConfig
+
+### OnnxConfigWithPast
+
+[[autodoc]] onnx.config.OnnxConfigWithPast
+
+### OnnxSeq2SeqConfigWithPast
+
+[[autodoc]] onnx.config.OnnxSeq2SeqConfigWithPast
+
+## ONNX Features
+
+每个ONNX配置与一组 _特性_ 相关联，使您能够为不同类型的拓扑结构或任务导出模型。
+
+### FeaturesManager
+
+[[autodoc]] onnx.features.FeaturesManager
+

docs/source/zh/main_classes/optimizer_schedules.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Optimization
+
+`.optimization` 模块提供了：
+
+- 一个带有固定权重衰减的优化器，可用于微调模型
+- 继承自 `_LRSchedule` 多个调度器：
+- 一个梯度累积类，用于累积多个批次的梯度
+
+## AdamW (PyTorch)
+
+[[autodoc]] AdamW
+
+## AdaFactor (PyTorch)
+
+[[autodoc]] Adafactor
+
+## AdamWeightDecay (TensorFlow)
+
+[[autodoc]] AdamWeightDecay
+
+[[autodoc]] create_optimizer
+
+## Schedules
+
+### Learning Rate Schedules (Pytorch)
+
+[[autodoc]] SchedulerType
+
+[[autodoc]] get_scheduler
+
+[[autodoc]] get_constant_schedule
+
+[[autodoc]] get_constant_schedule_with_warmup
+
+<img alt="" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/warmup_constant_schedule.png"/>
+
+[[autodoc]] get_cosine_schedule_with_warmup
+
+<img alt="" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/warmup_cosine_schedule.png"/>
+
+[[autodoc]] get_cosine_with_hard_restarts_schedule_with_warmup
+
+<img alt="" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/warmup_cosine_hard_restarts_schedule.png"/>
+
+[[autodoc]] get_linear_schedule_with_warmup
+
+<img alt="" src="https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/warmup_linear_schedule.png"/>
+
+[[autodoc]] get_polynomial_decay_schedule_with_warmup
+
+[[autodoc]] get_inverse_sqrt_schedule
+
+### Warmup (TensorFlow)
+
+[[autodoc]] WarmUp
+
+## Gradient Strategies
+
+### GradientAccumulator (TensorFlow)
+
+[[autodoc]] GradientAccumulator

docs/source/zh/main_classes/output.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# 模型输出
+
+所有模型的输出都是 [`~utils.ModelOutput`] 的子类的实例。这些是包含模型返回的所有信息的数据结构，但也可以用作元组或字典。
+
+让我们看一个例子：
+
+```python
+from transformers import BertTokenizer, BertForSequenceClassification
+import torch
+
+tokenizer = BertTokenizer.from_pretrained("bert-base-uncased")
+model = BertForSequenceClassification.from_pretrained("bert-base-uncased")
+
+inputs = tokenizer("Hello, my dog is cute", return_tensors="pt")
+labels = torch.tensor([1]).unsqueeze(0)  # Batch size 1
+outputs = model(**inputs, labels=labels)
+```
+
+`outputs` 对象是 [`~modeling_outputs.SequenceClassifierOutput`]，如下面该类的文档中所示，它表示它有一个可选的 `loss`，一个 `logits`，一个可选的 `hidden_states` 和一个可选的 `attentions` 属性。在这里，我们有 `loss`，因为我们传递了 `labels`，但我们没有 `hidden_states` 和 `attentions`，因为我们没有传递 `output_hidden_states=True` 或 `output_attentions=True`。
+
+<Tip>
+
+当传递 `output_hidden_states=True` 时，您可能希望 `outputs.hidden_states[-1]` 与 `outputs.last_hidden_states` 完全匹配。然而，这并不总是成立。一些模型在返回最后的 hidden state时对其应用归一化或其他后续处理。
+
+</Tip>
+
+
+您可以像往常一样访问每个属性，如果模型未返回该属性，您将得到 `None`。在这里，例如，`outputs.loss` 是模型计算的损失，而 `outputs.attentions` 是 `None`。
+
+当将我们的 `outputs` 对象视为元组时，它仅考虑那些没有 `None` 值的属性。例如这里它有两个元素，`loss` 和 `logits`，所以
+
+```python
+outputs[:2]
+```
+
+将返回元组 `(outputs.loss, outputs.logits)`。
+
+将我们的 `outputs` 对象视为字典时，它仅考虑那些没有 `None` 值的属性。例如在这里它有两个键，分别是 `loss` 和 `logits`。
+
+我们在这里记录了被多个类型模型使用的通用模型输出。特定输出类型在其相应的模型页面上有文档。
+
+## ModelOutput
+
+[[autodoc]] utils.ModelOutput
+    - to_tuple
+
+## BaseModelOutput
+
+[[autodoc]] modeling_outputs.BaseModelOutput
+
+## BaseModelOutputWithPooling
+
+[[autodoc]] modeling_outputs.BaseModelOutputWithPooling
+
+## BaseModelOutputWithCrossAttentions
+
+[[autodoc]] modeling_outputs.BaseModelOutputWithCrossAttentions
+
+## BaseModelOutputWithPoolingAndCrossAttentions
+
+[[autodoc]] modeling_outputs.BaseModelOutputWithPoolingAndCrossAttentions
+
+## BaseModelOutputWithPast
+
+[[autodoc]] modeling_outputs.BaseModelOutputWithPast
+
+## BaseModelOutputWithPastAndCrossAttentions
+
+[[autodoc]] modeling_outputs.BaseModelOutputWithPastAndCrossAttentions
+
+## Seq2SeqModelOutput
+
+[[autodoc]] modeling_outputs.Seq2SeqModelOutput
+
+## CausalLMOutput
+
+[[autodoc]] modeling_outputs.CausalLMOutput
+
+## CausalLMOutputWithCrossAttentions
+
+[[autodoc]] modeling_outputs.CausalLMOutputWithCrossAttentions
+
+## CausalLMOutputWithPast
+
+[[autodoc]] modeling_outputs.CausalLMOutputWithPast
+
+## MaskedLMOutput
+
+[[autodoc]] modeling_outputs.MaskedLMOutput
+
+## Seq2SeqLMOutput
+
+[[autodoc]] modeling_outputs.Seq2SeqLMOutput
+
+## NextSentencePredictorOutput
+
+[[autodoc]] modeling_outputs.NextSentencePredictorOutput
+
+## SequenceClassifierOutput
+
+[[autodoc]] modeling_outputs.SequenceClassifierOutput
+
+## Seq2SeqSequenceClassifierOutput
+
+[[autodoc]] modeling_outputs.Seq2SeqSequenceClassifierOutput
+
+## MultipleChoiceModelOutput
+
+[[autodoc]] modeling_outputs.MultipleChoiceModelOutput
+
+## TokenClassifierOutput
+
+[[autodoc]] modeling_outputs.TokenClassifierOutput
+
+## QuestionAnsweringModelOutput
+
+[[autodoc]] modeling_outputs.QuestionAnsweringModelOutput
+
+## Seq2SeqQuestionAnsweringModelOutput
+
+[[autodoc]] modeling_outputs.Seq2SeqQuestionAnsweringModelOutput
+
+## Seq2SeqSpectrogramOutput
+
+[[autodoc]] modeling_outputs.Seq2SeqSpectrogramOutput
+
+## SemanticSegmenterOutput
+
+[[autodoc]] modeling_outputs.SemanticSegmenterOutput
+
+## ImageClassifierOutput
+
+[[autodoc]] modeling_outputs.ImageClassifierOutput
+
+## ImageClassifierOutputWithNoAttention
+
+[[autodoc]] modeling_outputs.ImageClassifierOutputWithNoAttention
+
+## DepthEstimatorOutput
+
+[[autodoc]] modeling_outputs.DepthEstimatorOutput
+
+## Wav2Vec2BaseModelOutput
+
+[[autodoc]] modeling_outputs.Wav2Vec2BaseModelOutput
+
+## XVectorOutput
+
+[[autodoc]] modeling_outputs.XVectorOutput
+
+## Seq2SeqTSModelOutput
+
+[[autodoc]] modeling_outputs.Seq2SeqTSModelOutput
+
+## Seq2SeqTSPredictionOutput
+
+[[autodoc]] modeling_outputs.Seq2SeqTSPredictionOutput
+
+## SampleTSPredictionOutput
+
+[[autodoc]] modeling_outputs.SampleTSPredictionOutput
+
+## TFBaseModelOutput
+
+[[autodoc]] modeling_tf_outputs.TFBaseModelOutput
+
+## TFBaseModelOutputWithPooling
+
+[[autodoc]] modeling_tf_outputs.TFBaseModelOutputWithPooling
+
+## TFBaseModelOutputWithPoolingAndCrossAttentions
+
+[[autodoc]] modeling_tf_outputs.TFBaseModelOutputWithPoolingAndCrossAttentions
+
+## TFBaseModelOutputWithPast
+
+[[autodoc]] modeling_tf_outputs.TFBaseModelOutputWithPast
+
+## TFBaseModelOutputWithPastAndCrossAttentions
+
+[[autodoc]] modeling_tf_outputs.TFBaseModelOutputWithPastAndCrossAttentions
+
+## TFSeq2SeqModelOutput
+
+[[autodoc]] modeling_tf_outputs.TFSeq2SeqModelOutput
+
+## TFCausalLMOutput
+
+[[autodoc]] modeling_tf_outputs.TFCausalLMOutput
+
+## TFCausalLMOutputWithCrossAttentions
+
+[[autodoc]] modeling_tf_outputs.TFCausalLMOutputWithCrossAttentions
+
+## TFCausalLMOutputWithPast
+
+[[autodoc]] modeling_tf_outputs.TFCausalLMOutputWithPast
+
+## TFMaskedLMOutput
+
+[[autodoc]] modeling_tf_outputs.TFMaskedLMOutput
+
+## TFSeq2SeqLMOutput
+
+[[autodoc]] modeling_tf_outputs.TFSeq2SeqLMOutput
+
+## TFNextSentencePredictorOutput
+
+[[autodoc]] modeling_tf_outputs.TFNextSentencePredictorOutput
+
+## TFSequenceClassifierOutput
+
+[[autodoc]] modeling_tf_outputs.TFSequenceClassifierOutput
+
+## TFSeq2SeqSequenceClassifierOutput
+
+[[autodoc]] modeling_tf_outputs.TFSeq2SeqSequenceClassifierOutput
+
+## TFMultipleChoiceModelOutput
+
+[[autodoc]] modeling_tf_outputs.TFMultipleChoiceModelOutput
+
+## TFTokenClassifierOutput
+
+[[autodoc]] modeling_tf_outputs.TFTokenClassifierOutput
+
+## TFQuestionAnsweringModelOutput
+
+[[autodoc]] modeling_tf_outputs.TFQuestionAnsweringModelOutput
+
+## TFSeq2SeqQuestionAnsweringModelOutput
+
+[[autodoc]] modeling_tf_outputs.TFSeq2SeqQuestionAnsweringModelOutput
+
+## FlaxBaseModelOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxBaseModelOutput
+
+## FlaxBaseModelOutputWithPast
+
+[[autodoc]] modeling_flax_outputs.FlaxBaseModelOutputWithPast
+
+## FlaxBaseModelOutputWithPooling
+
+[[autodoc]] modeling_flax_outputs.FlaxBaseModelOutputWithPooling
+
+## FlaxBaseModelOutputWithPastAndCrossAttentions
+
+[[autodoc]] modeling_flax_outputs.FlaxBaseModelOutputWithPastAndCrossAttentions
+
+## FlaxSeq2SeqModelOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxSeq2SeqModelOutput
+
+## FlaxCausalLMOutputWithCrossAttentions
+
+[[autodoc]] modeling_flax_outputs.FlaxCausalLMOutputWithCrossAttentions
+
+## FlaxMaskedLMOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxMaskedLMOutput
+
+## FlaxSeq2SeqLMOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxSeq2SeqLMOutput
+
+## FlaxNextSentencePredictorOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxNextSentencePredictorOutput
+
+## FlaxSequenceClassifierOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxSequenceClassifierOutput
+
+## FlaxSeq2SeqSequenceClassifierOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxSeq2SeqSequenceClassifierOutput
+
+## FlaxMultipleChoiceModelOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxMultipleChoiceModelOutput
+
+## FlaxTokenClassifierOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxTokenClassifierOutput
+
+## FlaxQuestionAnsweringModelOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxQuestionAnsweringModelOutput
+
+## FlaxSeq2SeqQuestionAnsweringModelOutput
+
+[[autodoc]] modeling_flax_outputs.FlaxSeq2SeqQuestionAnsweringModelOutput

docs/source/zh/main_classes/pipelines.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Pipelines
+
+pipelines是使用模型进行推理的一种简单方法。这些pipelines是抽象了库中大部分复杂代码的对象，提供了一个专用于多个任务的简单API，包括专名识别、掩码语言建模、情感分析、特征提取和问答等。请参阅[任务摘要](../task_summary)以获取使用示例。
+
+有两种pipelines抽象类需要注意：
+
+- [`pipeline`]，它是封装所有其他pipelines的最强大的对象。
+- 针对特定任务pipelines，适用于[音频](#audio)、[计算机视觉](#computer-vision)、[自然语言处理](#natural-language-processing)和[多模态](#multimodal)任务。
+
+## pipeline抽象类
+
+*pipeline*抽象类是对所有其他可用pipeline的封装。它可以像任何其他pipeline一样实例化，但进一步提供额外的便利性。
+
+简单调用一个项目：
+
+
+```python
+>>> pipe = pipeline("text-classification")
+>>> pipe("This restaurant is awesome")
+[{'label': 'POSITIVE', 'score': 0.9998743534088135}]
+```
+
+如果您想使用 [hub](https://huggingface.co) 上的特定模型，可以忽略任务，如果hub上的模型已经定义了该任务：
+
+```python
+>>> pipe = pipeline(model="roberta-large-mnli")
+>>> pipe("This restaurant is awesome")
+[{'label': 'NEUTRAL', 'score': 0.7313136458396912}]
+```
+
+要在多个项目上调用pipeline，可以使用*列表*调用它。
+
+```python
+>>> pipe = pipeline("text-classification")
+>>> pipe(["This restaurant is awesome", "This restaurant is awful"])
+[{'label': 'POSITIVE', 'score': 0.9998743534088135},
+ {'label': 'NEGATIVE', 'score': 0.9996669292449951}]
+```
+
+为了遍历整个数据集，建议直接使用 `dataset`。这意味着您不需要一次性分配整个数据集，也不需要自己进行批处理。这应该与GPU上的自定义循环一样快。如果不是，请随时提出issue。
+
+```python
+import datasets
+from transformers import pipeline
+from transformers.pipelines.pt_utils import KeyDataset
+from tqdm.auto import tqdm
+
+pipe = pipeline("automatic-speech-recognition", model="facebook/wav2vec2-base-960h", device=0)
+dataset = datasets.load_dataset("superb", name="asr", split="test")
+
+# KeyDataset (only *pt*) will simply return the item in the dict returned by the dataset item
+# as we're not interested in the *target* part of the dataset. For sentence pair use KeyPairDataset
+for out in tqdm(pipe(KeyDataset(dataset, "file"))):
+    print(out)
+    # {"text": "NUMBER TEN FRESH NELLY IS WAITING ON YOU GOOD NIGHT HUSBAND"}
+    # {"text": ....}
+    # ....
+```
+
+为了方便使用，也可以使用生成器：
+
+
+```python
+from transformers import pipeline
+
+pipe = pipeline("text-classification")
+
+
+def data():
+    while True:
+        # This could come from a dataset, a database, a queue or HTTP request
+        # in a server
+        # Caveat: because this is iterative, you cannot use `num_workers > 1` variable
+        # to use multiple threads to preprocess data. You can still have 1 thread that
+        # does the preprocessing while the main runs the big inference
+        yield "This is a test"
+
+
+for out in pipe(data()):
+    print(out)
+    # {"text": "NUMBER TEN FRESH NELLY IS WAITING ON YOU GOOD NIGHT HUSBAND"}
+    # {"text": ....}
+    # ....
+```
+
+[[autodoc]] pipeline
+
+## Pipeline batching
+
+所有pipeline都可以使用批处理。这将在pipeline使用其流处理功能时起作用（即传递列表或 `Dataset` 或 `generator` 时）。
+
+```python
+from transformers import pipeline
+from transformers.pipelines.pt_utils import KeyDataset
+import datasets
+
+dataset = datasets.load_dataset("imdb", name="plain_text", split="unsupervised")
+pipe = pipeline("text-classification", device=0)
+for out in pipe(KeyDataset(dataset, "text"), batch_size=8, truncation="only_first"):
+    print(out)
+    # [{'label': 'POSITIVE', 'score': 0.9998743534088135}]
+    # Exactly the same output as before, but the content are passed
+    # as batches to the model
+```
+
+<Tip warning={true}>
+
+然而，这并不自动意味着性能提升。它可能是一个10倍的加速或5倍的减速，具体取决于硬件、数据和实际使用的模型。
+
+主要是加速的示例：
+
+</Tip>
+
+```python
+from transformers import pipeline
+from torch.utils.data import Dataset
+from tqdm.auto import tqdm
+
+pipe = pipeline("text-classification", device=0)
+
+
+class MyDataset(Dataset):
+    def __len__(self):
+        return 5000
+
+    def __getitem__(self, i):
+        return "This is a test"
+
+
+dataset = MyDataset()
+
+for batch_size in [1, 8, 64, 256]:
+    print("-" * 30)
+    print(f"Streaming batch_size={batch_size}")
+    for out in tqdm(pipe(dataset, batch_size=batch_size), total=len(dataset)):
+        pass
+```
+
+```
+# On GTX 970
+------------------------------
+Streaming no batching
+100%|██████████████████████████████████████████████████████████████████████| 5000/5000 [00:26<00:00, 187.52it/s]
+------------------------------
+Streaming batch_size=8
+100%|█████████████████████████████████████████████████████████████████████| 5000/5000 [00:04<00:00, 1205.95it/s]
+------------------------------
+Streaming batch_size=64
+100%|█████████████████████████████████████████████████████████████████████| 5000/5000 [00:02<00:00, 2478.24it/s]
+------------------------------
+Streaming batch_size=256
+100%|█████████████████████████████████████████████████████████████████████| 5000/5000 [00:01<00:00, 2554.43it/s]
+(diminishing returns, saturated the GPU)
+```
+
+主要是减速的示例：
+
+```python
+class MyDataset(Dataset):
+    def __len__(self):
+        return 5000
+
+    def __getitem__(self, i):
+        if i % 64 == 0:
+            n = 100
+        else:
+            n = 1
+        return "This is a test" * n
+```
+
+与其他句子相比，这是一个非常长的句子。在这种情况下，**整个**批次将需要400个tokens的长度，因此整个批次将是 [64, 400] 而不是 [64, 4]，从而导致较大的减速。更糟糕的是，在更大的批次上，程序会崩溃。
+
+```
+------------------------------
+Streaming no batching
+100%|█████████████████████████████████████████████████████████████████████| 1000/1000 [00:05<00:00, 183.69it/s]
+------------------------------
+Streaming batch_size=8
+100%|█████████████████████████████████████████████████████████████████████| 1000/1000 [00:03<00:00, 265.74it/s]
+------------------------------
+Streaming batch_size=64
+100%|██████████████████████████████████████████████████████████████████████| 1000/1000 [00:26<00:00, 37.80it/s]
+------------------------------
+Streaming batch_size=256
+  0%|                                                                                 | 0/1000 [00:00<?, ?it/s]
+Traceback (most recent call last):
+  File "/home/nicolas/src/transformers/test.py", line 42, in <module>
+    for out in tqdm(pipe(dataset, batch_size=256), total=len(dataset)):
+....
+    q = q / math.sqrt(dim_per_head)  # (bs, n_heads, q_length, dim_per_head)
+RuntimeError: CUDA out of memory. Tried to allocate 376.00 MiB (GPU 0; 3.95 GiB total capacity; 1.72 GiB already allocated; 354.88 MiB free; 2.46 GiB reserved in total by PyTorch)
+```
+
+对于这个问题，没有好的（通用）解决方案，效果可能因您的用例而异。经验法则如下：
+
+对于用户，一个经验法则是：
+
+- **使用硬件测量负载性能。测量、测量、再测量。真实的数字是唯一的方法。**
+- 如果受到延迟的限制（进行推理的实时产品），不要进行批处理。
+- 如果使用CPU，不要进行批处理。
+- 如果您在GPU上处理的是吞吐量（您希望在大量静态数据上运行模型），则：
+  - 如果对序列长度的大小没有概念（"自然"数据），默认情况下不要进行批处理，进行测试并尝试逐渐添加，添加OOM检查以在失败时恢复（如果您不能控制序列长度，它将在某些时候失败）。
+  - 如果您的序列长度非常规律，那么批处理更有可能非常有趣，进行测试并推动它，直到出现OOM。
+  - GPU越大，批处理越有可能变得更有趣
+- 一旦启用批处理，确保能够很好地处理OOM。
+
+## Pipeline chunk batching
+
+`zero-shot-classification` 和 `question-answering` 在某种意义上稍微特殊，因为单个输入可能会导致模型的多次前向传递。在正常情况下，这将导致 `batch_size` 参数的问题。
+
+为了规避这个问题，这两个pipeline都有点特殊，它们是 `ChunkPipeline` 而不是常规的 `Pipeline`。简而言之：
+
+
+```python
+preprocessed = pipe.preprocess(inputs)
+model_outputs = pipe.forward(preprocessed)
+outputs = pipe.postprocess(model_outputs)
+```
+
+现在变成：
+
+
+```python
+all_model_outputs = []
+for preprocessed in pipe.preprocess(inputs):
+    model_outputs = pipe.forward(preprocessed)
+    all_model_outputs.append(model_outputs)
+outputs = pipe.postprocess(all_model_outputs)
+```
+
+这对您的代码应该是非常直观的，因为pipeline的使用方式是相同的。
+
+这是一个简化的视图，因为Pipeline可以自动处理批次！这意味着您不必担心您的输入实际上会触发多少次前向传递，您可以独立于输入优化 `batch_size`。前面部分的注意事项仍然适用。
+
+## Pipeline自定义
+
+如果您想要重载特定的pipeline。
+
+请随时为您手头的任务创建一个issue，Pipeline的目标是易于使用并支持大多数情况，因此 `transformers` 可能支持您的用例。
+
+如果您想简单地尝试一下，可以：
+
+- 继承您选择的pipeline
+
+```python
+class MyPipeline(TextClassificationPipeline):
+    def postprocess():
+        # Your code goes here
+        scores = scores * 100
+        # And here
+
+
+my_pipeline = MyPipeline(model=model, tokenizer=tokenizer, ...)
+# or if you use *pipeline* function, then:
+my_pipeline = pipeline(model="xxxx", pipeline_class=MyPipeline)
+```
+
+这样就可以让您编写所有想要的自定义代码。
+
+
+## 实现一个pipeline
+
+[实现一个新的pipeline](../add_new_pipeline)
+
+## 音频
+
+可用于音频任务的pipeline包括以下几种。
+
+### AudioClassificationPipeline
+
+[[autodoc]] AudioClassificationPipeline
+    - __call__
+    - all
+
+### AutomaticSpeechRecognitionPipeline
+
+[[autodoc]] AutomaticSpeechRecognitionPipeline
+    - __call__
+    - all
+
+### TextToAudioPipeline
+
+[[autodoc]] TextToAudioPipeline
+    - __call__
+    - all
+
+
+### ZeroShotAudioClassificationPipeline
+
+[[autodoc]] ZeroShotAudioClassificationPipeline
+    - __call__
+    - all
+
+## 计算机视觉
+
+可用于计算机视觉任务的pipeline包括以下几种。
+
+### DepthEstimationPipeline
+[[autodoc]] DepthEstimationPipeline
+    - __call__
+    - all
+
+### ImageClassificationPipeline
+
+[[autodoc]] ImageClassificationPipeline
+    - __call__
+    - all
+
+### ImageSegmentationPipeline
+
+[[autodoc]] ImageSegmentationPipeline
+    - __call__
+    - all
+
+### ImageToImagePipeline
+
+[[autodoc]] ImageToImagePipeline
+    - __call__
+    - all
+
+### ObjectDetectionPipeline
+
+[[autodoc]] ObjectDetectionPipeline
+    - __call__
+    - all
+
+### VideoClassificationPipeline
+
+[[autodoc]] VideoClassificationPipeline
+    - __call__
+    - all
+
+### ZeroShotImageClassificationPipeline
+
+[[autodoc]] ZeroShotImageClassificationPipeline
+    - __call__
+    - all
+
+### ZeroShotObjectDetectionPipeline
+
+[[autodoc]] ZeroShotObjectDetectionPipeline
+    - __call__
+    - all
+
+## 自然语言处理
+
+可用于自然语言处理任务的pipeline包括以下几种。
+
+### ConversationalPipeline
+
+[[autodoc]] Conversation
+
+[[autodoc]] ConversationalPipeline
+    - __call__
+    - all
+
+### FillMaskPipeline
+
+[[autodoc]] FillMaskPipeline
+    - __call__
+    - all
+
+### NerPipeline
+
+[[autodoc]] NerPipeline
+
+See [`TokenClassificationPipeline`] for all details.
+
+### QuestionAnsweringPipeline
+
+[[autodoc]] QuestionAnsweringPipeline
+    - __call__
+    - all
+
+### SummarizationPipeline
+
+[[autodoc]] SummarizationPipeline
+    - __call__
+    - all
+
+### TableQuestionAnsweringPipeline
+
+[[autodoc]] TableQuestionAnsweringPipeline
+    - __call__
+
+### TextClassificationPipeline
+
+[[autodoc]] TextClassificationPipeline
+    - __call__
+    - all
+
+### TextGenerationPipeline
+
+[[autodoc]] TextGenerationPipeline
+    - __call__
+    - all
+
+### Text2TextGenerationPipeline
+
+[[autodoc]] Text2TextGenerationPipeline
+    - __call__
+    - all
+
+### TokenClassificationPipeline
+
+[[autodoc]] TokenClassificationPipeline
+    - __call__
+    - all
+
+### TranslationPipeline
+
+[[autodoc]] TranslationPipeline
+    - __call__
+    - all
+
+### ZeroShotClassificationPipeline
+
+[[autodoc]] ZeroShotClassificationPipeline
+    - __call__
+    - all
+
+## 多模态 
+
+可用于多模态任务的pipeline包括以下几种。
+
+### DocumentQuestionAnsweringPipeline
+
+[[autodoc]] DocumentQuestionAnsweringPipeline
+    - __call__
+    - all
+
+### FeatureExtractionPipeline
+
+[[autodoc]] FeatureExtractionPipeline
+    - __call__
+    - all
+
+### ImageToTextPipeline
+
+[[autodoc]] ImageToTextPipeline
+    - __call__
+    - all
+
+### MaskGenerationPipeline
+
+[[autodoc]] MaskGenerationPipeline
+    - __call__
+    - all
+
+### VisualQuestionAnsweringPipeline
+
+[[autodoc]] VisualQuestionAnsweringPipeline
+    - __call__
+    - all
+
+## Parent class: `Pipeline`
+
+[[autodoc]] Pipeline

docs/source/zh/main_classes/processors.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Processors
+
+在 Transformers 库中，processors可以有两种不同的含义：
+- 为多模态模型，例如[Wav2Vec2](../model_doc/wav2vec2)（语音和文本）或[CLIP](../model_doc/clip)（文本和视觉）预处理输入的对象
+- 在库的旧版本中用于预处理GLUE或SQUAD数据的已弃用对象。
+
+## 多模态processors
+
+任何多模态模型都需要一个对象来编码或解码将多个模态（包括文本、视觉和音频）组合在一起的数据。这由称为processors的对象处理，这些processors将两个或多个处理对象组合在一起，例如tokenizers（用于文本模态），image processors（用于视觉）和feature extractors（用于音频）。
+
+这些processors继承自以下实现保存和加载功能的基类：
+
+
+[[autodoc]] ProcessorMixin
+
+## 已弃用的processors
+
+所有processor都遵循与 [`~data.processors.utils.DataProcessor`] 相同的架构。processor返回一个 [`~data.processors.utils.InputExample`] 列表。这些 [`~data.processors.utils.InputExample`] 可以转换为 [`~data.processors.utils.InputFeatures`] 以供输送到模型。
+
+[[autodoc]] data.processors.utils.DataProcessor
+
+[[autodoc]] data.processors.utils.InputExample
+
+[[autodoc]] data.processors.utils.InputFeatures
+
+## GLUE
+
+[General Language Understanding Evaluation (GLUE)](https://gluebenchmark.com/) 是一个基准测试，评估模型在各种现有的自然语言理解任务上的性能。它与论文 [GLUE: A multi-task benchmark and analysis platform for natural language understanding](https://openreview.net/pdf?id=rJ4km2R5t7) 一同发布。
+
+该库为以下任务提供了总共10个processor：MRPC、MNLI、MNLI（mismatched）、CoLA、SST2、STSB、QQP、QNLI、RTE 和 WNLI。
+
+这些processor是：
+
+- [`~data.processors.utils.MrpcProcessor`]
+- [`~data.processors.utils.MnliProcessor`]
+- [`~data.processors.utils.MnliMismatchedProcessor`]
+- [`~data.processors.utils.Sst2Processor`]
+- [`~data.processors.utils.StsbProcessor`]
+- [`~data.processors.utils.QqpProcessor`]
+- [`~data.processors.utils.QnliProcessor`]
+- [`~data.processors.utils.RteProcessor`]
+- [`~data.processors.utils.WnliProcessor`]
+
+此外，还可以使用以下方法从数据文件加载值并将其转换为 [`~data.processors.utils.InputExample`] 列表。
+
+[[autodoc]] data.processors.glue.glue_convert_examples_to_features
+
+
+## XNLI
+
+[跨语言NLI语料库（XNLI）](https://www.nyu.edu/projects/bowman/xnli/) 是一个评估跨语言文本表示质量的基准测试。XNLI是一个基于[*MultiNLI*](http://www.nyu.edu/projects/bowman/multinli/)的众包数据集：”文本对“被标记为包含15种不同语言（包括英语等高资源语言和斯瓦希里语等低资源语言）的文本蕴涵注释。
+
+它与论文 [XNLI: Evaluating Cross-lingual Sentence Representations](https://arxiv.org/abs/1809.05053) 一同发布。
+
+该库提供了加载XNLI数据的processor：
+
+- [`~data.processors.utils.XnliProcessor`]
+
+请注意，由于测试集上有“gold”标签，因此评估是在测试集上进行的。
+
+使用这些processor的示例在 [run_xnli.py](https://github.com/huggingface/transformers/tree/main/examples/pytorch/text-classification/run_xnli.py) 脚本中提供。
+
+
+## SQuAD
+
+[斯坦福问答数据集（SQuAD）](https://rajpurkar.github.io/SQuAD-explorer//) 是一个评估模型在问答上性能的基准测试。有两个版本，v1.1 和 v2.0。第一个版本（v1.1）与论文 [SQuAD: 100,000+ Questions for Machine Comprehension of Text](https://arxiv.org/abs/1606.05250) 一同发布。第二个版本（v2.0）与论文 [Know What You Don't Know: Unanswerable Questions for SQuAD](https://arxiv.org/abs/1806.03822) 一同发布。
+
+该库为两个版本各自提供了一个processor：
+
+### Processors
+
+这两个processor是：
+
+- [`~data.processors.utils.SquadV1Processor`]
+- [`~data.processors.utils.SquadV2Processor`]
+
+它们都继承自抽象类 [`~data.processors.utils.SquadProcessor`]。
+
+[[autodoc]] data.processors.squad.SquadProcessor
+    - all
+
+此外，可以使用以下方法将 SQuAD 示例转换为可用作模型输入的 [`~data.processors.utils.SquadFeatures`]。
+
+[[autodoc]] data.processors.squad.squad_convert_examples_to_features
+
+
+这些processor以及前面提到的方法可以与包含数据的文件以及tensorflow_datasets包一起使用。下面给出了示例。
+
+
+### Example使用
+
+以下是使用processor以及使用数据文件的转换方法的示例：
+
+```python
+# Loading a V2 processor
+processor = SquadV2Processor()
+examples = processor.get_dev_examples(squad_v2_data_dir)
+
+# Loading a V1 processor
+processor = SquadV1Processor()
+examples = processor.get_dev_examples(squad_v1_data_dir)
+
+features = squad_convert_examples_to_features(
+    examples=examples,
+    tokenizer=tokenizer,
+    max_seq_length=max_seq_length,
+    doc_stride=args.doc_stride,
+    max_query_length=max_query_length,
+    is_training=not evaluate,
+)
+```
+
+使用 *tensorflow_datasets* 就像使用数据文件一样简单：
+
+```python
+# tensorflow_datasets only handle Squad V1.
+tfds_examples = tfds.load("squad")
+examples = SquadV1Processor().get_examples_from_dataset(tfds_examples, evaluate=evaluate)
+
+features = squad_convert_examples_to_features(
+    examples=examples,
+    tokenizer=tokenizer,
+    max_seq_length=max_seq_length,
+    doc_stride=args.doc_stride,
+    max_query_length=max_query_length,
+    is_training=not evaluate,
+)
+```
+
+另一个使用这些processor的示例在 [run_squad.py](https://github.com/huggingface/transformers/tree/main/examples/legacy/question-answering/run_squad.py) 脚本中提供。
\ No newline at end of file

docs/source/zh/main_classes/text_generation.md

+<!--Copyright 2022 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Generation
+
+每个框架都在它们各自的 `GenerationMixin` 类中实现了文本生成的 `generate` 方法：
+
+- PyTorch [`~generation.GenerationMixin.generate`] 在 [`~generation.GenerationMixin`] 中实现。
+- TensorFlow [`~generation.TFGenerationMixin.generate`] 在 [`~generation.TFGenerationMixin`] 中实现。
+- Flax/JAX [`~generation.FlaxGenerationMixin.generate`] 在 [`~generation.FlaxGenerationMixin`] 中实现。
+
+无论您选择哪个框架，都可以使用 [`~generation.GenerationConfig`] 类实例对 generate 方法进行参数化。有关生成方法的控制参数的完整列表，请参阅此类。
+
+要了解如何检查模型的生成配置、默认值是什么、如何临时更改参数以及如何创建和保存自定义生成配置，请参阅 [文本生成策略指南](../generation_strategies)。该指南还解释了如何使用相关功能，如token流。
+
+## GenerationConfig
+
+[[autodoc]] generation.GenerationConfig
+	- from_pretrained
+	- from_model_config
+	- save_pretrained
+
+## GenerationMixin
+
+[[autodoc]] generation.GenerationMixin
+	- generate
+	- compute_transition_scores
+	- greedy_search
+	- sample
+	- beam_search
+	- beam_sample
+	- contrastive_search
+	- group_beam_search
+	- constrained_beam_search
+
+## TFGenerationMixin
+
+[[autodoc]] generation.TFGenerationMixin
+	- generate
+	- compute_transition_scores
+
+## FlaxGenerationMixin
+
+[[autodoc]] generation.FlaxGenerationMixin
+	- generate

docs/source/zh/main_classes/tokenizer.md

+<!--Copyright 2020 The HuggingFace Team. All rights reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with
+the License. You may obtain a copy of the License at
+
+http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on
+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
+rendered properly in your Markdown viewer.
+
+-->
+
+# Tokenizer
+
+tokenizer负责准备输入以供模型使用。该库包含所有模型的tokenizer。大多数tokenizer都有两种版本：一个是完全的 Python 实现，另一个是基于 Rust 库 [🤗 Tokenizers](https://github.com/huggingface/tokenizers) 的“Fast”实现。"Fast" 实现允许：
+
+1. 在批量分词时显著提速
+2. 在原始字符串（字符和单词）和token空间之间进行映射的其他方法（例如，获取包含给定字符的token的索引或与给定token对应的字符范围）。
+
+基类 [PreTrainedTokenizer] 和 [PreTrained TokenizerFast] 实现了在模型输入中编码字符串输入的常用方法（见下文），并从本地文件或目录或从库提供的预训练的 tokenizer（从 HuggingFace 的 AWS S3 存储库下载）实例化/保存 python 和“Fast” tokenizer。它们都依赖于包含常用方法的 [`~tokenization_utils_base.PreTrainedTokenizerBase`]和[`~tokenization_utils_base.SpecialTokensMixin`]。
+
+因此，[`PreTrainedTokenizer`] 和 [`PreTrainedTokenizerFast`] 实现了使用所有tokenizers的主要方法：
+
+- 分词（将字符串拆分为子词标记字符串），将tokens字符串转换为id并转换回来，以及编码/解码（即标记化并转换为整数）。
+- 以独立于底层结构（BPE、SentencePiece……）的方式向词汇表中添加新tokens。
+- 管理特殊tokens（如mask、句首等）：添加它们，将它们分配给tokenizer中的属性以便于访问，并确保它们在标记过程中不会被分割。
+
+[`BatchEncoding`] 包含 [`~tokenization_utils_base.PreTrainedTokenizerBase`] 的编码方法（`__call__`、`encode_plus` 和 `batch_encode_plus`）的输出，并且是从 Python 字典派生的。当tokenizer是纯 Python tokenizer时，此类的行为就像标准的 Python 字典一样，并保存这些方法计算的各种模型输入（`input_ids`、`attention_mask` 等）。当分词器是“Fast”分词器时（即由 HuggingFace 的 [tokenizers 库](https://github.com/huggingface/tokenizers) 支持），此类还提供了几种高级对齐方法，可用于在原始字符串（字符和单词）与token空间之间进行映射（例如，获取包含给定字符的token的索引或与给定token对应的字符范围）。
+
+
+## PreTrainedTokenizer
+
+[[autodoc]] PreTrainedTokenizer
+    - __call__
+    - add_tokens
+    - add_special_tokens
+    - apply_chat_template
+    - batch_decode
+    - decode
+    - encode
+    - push_to_hub
+    - all
+
+## PreTrainedTokenizerFast
+
+[`PreTrainedTokenizerFast`] 依赖于 [tokenizers](https://huggingface.co/docs/tokenizers) 库。可以非常简单地将从 🤗 tokenizers 库获取的tokenizers加载到 🤗 transformers 中。查看 [使用 🤗 tokenizers 的分词器](../fast_tokenizers) 页面以了解如何执行此操作。
+
+[[autodoc]] PreTrainedTokenizerFast
+    - __call__
+    - add_tokens
+    - add_special_tokens
+    - apply_chat_template
+    - batch_decode
+    - decode
+    - encode
+    - push_to_hub
+    - all
+
+## BatchEncoding
+
+[[autodoc]] BatchEncoding