diff --git a/_typos.toml b/_typos.toml
index fb4098ec8f2..1faa0f64f36 100644
--- a/_typos.toml
+++ b/_typos.toml
@@ -36,10 +36,6 @@ Similarily = "Similarily"
 Simle = "Simle"
 Sovler = "Sovler"
 Successed = "Successed"
-classfy = "classfy"
-corss = "corss"
-creatation = "creatation"
-creats = "creats"
 dafault = "dafault"
 datas = "datas"
 decribe = "decribe"
diff --git a/docs/design/motivation/api.md b/docs/design/motivation/api.md
index 87eca5bd72a..b2f407f78a4 100644
--- a/docs/design/motivation/api.md
+++ b/docs/design/motivation/api.md
@@ -61,7 +61,7 @@ fQ = f(paddle.layer.data(input_name="Q"))
 
 topology = paddle.layer.less_than(
                paddle.layer.cross_entropy(fA, fQ),
-               paddle.layer.corss_entropy(fB, fQ))
+               paddle.layer.cross_entropy(fB, fQ))
 
 # Derive parameters required in topology and create them in model.
 parameters = paddle.parameters.create(topology)
diff --git a/docs/design/others/simple_op_design.md b/docs/design/others/simple_op_design.md
index c93c21affe9..13c1ec35b0b 100644
--- a/docs/design/others/simple_op_design.md
+++ b/docs/design/others/simple_op_design.md
@@ -64,7 +64,7 @@ message OpProto {
 To generate Python code automatically:
 
 ```python
-def create_python_ops_creatation_functions():
+def create_python_ops_creation_functions():
     op_protos = paddle.framework.OpRegistry.get_all_op_proto()
     for type_name in op_protos:
         op_proto = op_protos[type_name]
@@ -77,7 +77,7 @@ def create_python_ops_creatation_functions():
         __impl__.__doc__ = create_doc_string(op_proto)
         globals()[type_name] = __impl__
 
-create_python_ops_creatation_functions()
+create_python_ops_creation_functions()
 ```
 
 ### Message from Python to C++
diff --git a/docs/dev_guides/custom_device_docs/stream_api_en.md b/docs/dev_guides/custom_device_docs/stream_api_en.md
index 9c854052773..eea401432fa 100644
--- a/docs/dev_guides/custom_device_docs/stream_api_en.md
+++ b/docs/dev_guides/custom_device_docs/stream_api_en.md
@@ -10,7 +10,7 @@ C_Status (*create_stream)(const C_Device device, C_Stream* stream)
 
 ### Description
 
-It creats a stream, which is used to execute asynchronous tasks within the framework. In the stream, tasks are done in order. When the device does not support asynchronous execution, the API is required to be implemented with an empty method.
+It creates a stream, which is used to execute asynchronous tasks within the framework. In the stream, tasks are done in order. When the device does not support asynchronous execution, the API is required to be implemented with an empty method.
 
 ### Parameter
 
diff --git a/docs/guides/model_convert/convert_from_pytorch/nlp_migration_experiences_cn.md b/docs/guides/model_convert/convert_from_pytorch/nlp_migration_experiences_cn.md
index a9de43df157..1642f47c404 100644
--- a/docs/guides/model_convert/convert_from_pytorch/nlp_migration_experiences_cn.md
+++ b/docs/guides/model_convert/convert_from_pytorch/nlp_migration_experiences_cn.md
@@ -354,7 +354,7 @@ A：有的，例如：
 下面以 BERT 模型为例介绍转换的过程：
 
 - 如果 PyTorch repo 中已经提供权重，那么可以直接下载并进行后续的转换。huggingface 的 transformers 中提供了大部分模型参数，使用模型权重名称`model_name_or_path`即可加载（如`bert-base-uncased`）。或者从 huggingface 官网直接下载：https://huggingface.co/bert-base-uncased/tree/main
-- 如果没有提供，则可以基于 PyTorch 代码，随机生成一个初始化权重(定义完 model 以后，使用`torch.save()` API 保存模型权重)，然后将生成的 PyTorch 权重 (`bert_sequence_classfy.pth`) 转换为飞桨模型权重。
+- 如果没有提供，则可以基于 PyTorch 代码，随机生成一个初始化权重(定义完 model 以后，使用`torch.save()` API 保存模型权重)，然后将生成的 PyTorch 权重 (`bert_sequence_classify.pth`) 转换为飞桨模型权重。
 
 ```python
 from transformers import BertModel
@@ -366,7 +366,7 @@ PATH = './torch_weight.bin'
 torch.save(hf_model.state_dict(), PATH)
 ```
 
-然后将生成的 PyTorch 权重 (`bert_sequence_classfy.pth`) 转换为飞桨模型权重，转换代码如下（代码解释详见代码后的 FAQ）
+然后将生成的 PyTorch 权重 (`bert_sequence_classify.pth`) 转换为飞桨模型权重，转换代码如下（代码解释详见代码后的 FAQ）
 
 ```python
 def convert_pytorch_checkpoint_to_paddle(