typos修复

Author: dai-junjie

_typos.toml

@@ -31,16 +31,9 @@ Classfication = "Classfication"
 Comparision = "Comparision"
 Contructing = "Contructing"
 Creenshot = "Creenshot"
-DELCARE = "DELCARE"
-Dateset = "Dateset"
-Discription = "Discription"
-Distrbuted = "Distrbuted"
-Driect = "Driect"
 Embeddding = "Embeddding"
 Embeding = "Embeding"
 Engish = "Engish"
-Fasle = "Fasle"
-Flase = "Flase"
 Generater = "Generater"
 Gloabal = "Gloabal"
 Imporvement = "Imporvement"

docs/api/paddle/static/IpuStrategy_cn.rst

@@ -46,7 +46,7 @@ set_pipelining_config(self, enable_pipelining, batches_per_step, enable_gradient
 
     - **enable_pipelining** (bool，可选)- 是否使能子图之间的数据流水线。仅支持当 enable_manual_shard=True 时，enable_pipelining 可以置为 True。默认值为 False，表示不使能该功能。
     - **batches_per_step** (int，可选)- 指定数据流水线每次运算多少个 batch 的数据。默认值为 1，表示不使能数据流水线功能。
-    - **enable_gradient_accumulation** (bool，可选)- 是否使能梯度累积，只用于训练模式。默认值为 Flase，表示不使能梯度累积功能。
+    - **enable_gradient_accumulation** (bool，可选)- 是否使能梯度累积，只用于训练模式。默认值为 False，表示不使能梯度累积功能。
     - **accumulation_factor** (int，可选)- 指定累积运算多少个 batch 更新一次权重。默认值为 1，表示不使能权重累积更新功能。
 
 **代码示例**

docs/design/dist_train/mpi_enabled_design.md

@@ -8,7 +8,7 @@ When we do distribute multi GPU training, the communication overhead between ser
 
 We will use OpenMPI API to PaddlePaddle, which can bring two benefits to PaddlePaddle:
 1. Enable RDMA with PaddlePaddle, which bring high-performance low latency networks.
-2. Enable GPUDriect with PaddlePaddle, which bring the highest throughput and lowest latency GPU read and write.
+2. Enable GPUDirect with PaddlePaddle, which bring the highest throughput and lowest latency GPU read and write.
 
 # Change list
 * Compile args: Need add compile args to enable MPI support.

docs/design/phi/design_cn.md

@@ -1134,7 +1134,7 @@ class CompatMetaTensor : public phi::MetaTensor {
 通过前面介绍的 PT_INFER_META 宏归一化函数形式，然后将`PT_INFER_META(***InferMeta)`包装到一个 functor 中，functor 中先将 InferShapeContext 转换为 InferMetaContext，再调用相应 InferMeta 函数，通过一个宏统一管理代码
 
 ```
-##define DELCARE_INFER_SHAPE_FUNCTOR(op_type, functor_name, fn)      \
+##define DECLARE_INFER_SHAPE_FUNCTOR(op_type, functor_name, fn)      \
   struct functor_name : public paddle::framework::InferShapeBase {  \
     void operator()(                                                \
         paddle::framework::InferShapeContext* ctx) const override { \
@@ -1176,7 +1176,7 @@ class SignOp : public framework::OperatorWithKernel {
   using framework::OperatorWithKernel::OperatorWithKernel;
 };
 
-DELCARE_INFER_SHAPE_FUNCTOR(
+DECLARE_INFER_SHAPE_FUNCTOR(
     sign, SignInferShapeFunctor, PT_INFER_META(phi::UnchangedInferMetaNew));
 REGISTER_OPERATOR(sign, ops::SignOp, ops::SignOpMaker<float>,
                   ops::SignGradMaker<paddle::framework::OpDesc>,

docs/design/phi/design_en.md

@@ -1134,7 +1134,7 @@ Then, in order to transplant the functional `InferMeta` back to the Op architect
 Normalize the function form through the `PT_INFER_META` macro, and then wrap `PT_INFER_META(***InferMeta)` into a functor. The functor first converts the `InferShapeContext` to `InferMetaContext`, then calls the corresponding `InferMeta` function, and manages the code uniformly through a macro.
 
 ```
-##define DELCARE_INFER_SHAPE_FUNCTOR(op_type, functor_name, fn)      \
+##define DECLARE_INFER_SHAPE_FUNCTOR(op_type, functor_name, fn)      \
   struct functor_name : public paddle::framework::InferShapeBase {  \
     void operator()(                                                \
         paddle::framework::InferShapeContext* ctx) const override { \
@@ -1176,7 +1176,7 @@ class SignOp : public framework::OperatorWithKernel {
   using framework::OperatorWithKernel::OperatorWithKernel;
 };
 
-DELCARE_INFER_SHAPE_FUNCTOR(
+DECLARE_INFER_SHAPE_FUNCTOR(
     sign, SignInferShapeFunctor, PT_INFER_META(phi::UnchangedInferMetaNew));
 REGISTER_OPERATOR(sign, ops::SignOp, ops::SignOpMaker<float>,
                   ops::SignGradMaker<paddle::framework::OpDesc>,

docs/eval/evaluation_of_docs_system.md

@@ -549,7 +549,7 @@ MindSpore 的有自己独立的文档分类标准和风格，所以硬套本文
 | 数据加载与预处理             | Images CSV Numpy pandas.DataFrame TFRecord and tf.Example Additional formats with tf.io Text More text loading Classifying structured data with preprocessing layers Classfication on imbalanced data Time series forecasting Decision forest models | 13   | Datasets & Dataloaders                                       | 1    | 数据处理 数据处理（进阶） 自动数据增强 轻量化数据处理 单节点数据缓存 优化数据处理 | 6    | 数据集的定义和加载 数据预处理                                | 2      |
 | 如何组网                     | Modules, layers, and models                                  | 1    | Build the Neural Network Building Models with PyTorch What is torch.nn really? Learing PyTorch with Examples | 4    | 创建网络 网络构建                                            | 2    | 模型组网 飞桨高层 API 使用指南 层与模型                      | 3      |
 | 如何训练                     | Training loops NumPy API Checkpoint SavedModel               | 4    | Optimization Model Parameters Traning wiht PyTorch           | 2    | 模型训练 训练与评估                                          | 2    | 训练与预测验证 自定义指标                                    | 2      |
-| 保存与加载模型               | Save and load Save and load(Distrbuted Training)             | 2    | Save and Load the Model                                      | 1    | 保存与加载                                                   | 1    | 模型保存与载入 模型保存及加载（应用实践）                    | 2      |
+| 保存与加载模型               | Save and load Save and load(Distributed Training)             | 2    | Save and Load the Model                                      | 1    | 保存与加载                                                   | 1    | 模型保存与载入 模型保存及加载（应用实践）                    | 2      |
 | 可视化、调优技巧             | Overfit and underfit Tune hyperprameters with Keras Tuner Better performance with tf.function Profile TensorFlow performance Graph optimizaition Optimize GPU Performance Mixed precision | 7    | PyTorch TensorBoard Support Model Understanding with Captum Visualizing Models, Data, and Training with TensorBoard Profiling your PyTorch Module PyTorch Profiler with TensorBoard Hyperparameter tuning with Ray Tune Optimizing Vision Transformer Model for Deployment Parametrization Tutorial Pruning Tutorial Grokking PyTorch Intel CPU performance from first principles | 11   | 查看中间文件 Dump 功能调试 自定义调试信息 调用自定义类 算子增量编译 算子调优工具 自动数据加速 固定随机性以复现脚本运行结果 | 8    | VisualDL 工具简介 VisualDL 使用指南 飞桨模型量化             | 3      |
 | 自动微分                     | Automatic differentiation Advanced autodiff                  | 2    | Automatic Differentiation with torch.autograd The Fundamentals of Autograd | 2    | 自动微分                                                     | 1    | 自动微分                                                     | 1      |
 | 动态图与静态图               | Graphs and functions                                         | 1    | （torchscript 其实是静态图，不过归类到部署中了）             | 0    | 动态图与静态图                                               | 1    | 使用样例 转换原理 支持语法 案例解析 报错调试 动态图 使用动转静完成以图搜图 | 7      |

docs/guides/model_convert/convert_from_pytorch/api_difference/paddle_more_args/torch.utils.data.BatchSampler.md

@@ -10,7 +10,7 @@ torch.utils.data.BatchSampler(sampler,
 ```python
 paddle.io.BatchSampler(dataset=None,
                        sampler=None,
-                       shuffle=Fasle,
+                       shuffle=False,
                        batch_size=1,
                        drop_last=False)
 ```

docs/install/Tables_en.md

@@ -277,7 +277,7 @@ PaddePaddle implements references to various BLAS/CUDA/cuDNN libraries by specif
     <thead>
     <tr>
         <th> Version Number </th>
-        <th> Release Discription </th>
+        <th> Release Description </th>
     </tr>
     </thead>
     <tbody>