From a1ba7e578ce4b34e80fadfc480e87e92b3ebf93b Mon Sep 17 00:00:00 2001
From: Chijunlong <2869897445@qq.com>
Date: Tue, 28 Dec 2021 11:01:04 +0800
Subject: [PATCH] update mindtext.embeddings/index.rst and installation.rst

add mindtext.dataset.test1.md
---
 .../mindtext.dataset.test1.md                 | 56 ++++++++++++++++++-
 .../source/apis/mindtext.embeddings/index.rst |  4 --
 docs/source/user/installation.rst             |  9 +--
 3 files changed, 60 insertions(+), 9 deletions(-)

diff --git a/docs/source/apis/mindtext.dataset/mindtext.dataset.test1.md b/docs/source/apis/mindtext.dataset/mindtext.dataset.test1.md
index 2cb259c..c886cd1 100644
--- a/docs/source/apis/mindtext.dataset/mindtext.dataset.test1.md
+++ b/docs/source/apis/mindtext.dataset/mindtext.dataset.test1.md
@@ -1 +1,55 @@
-# mindtext.dataset.test1
\ No newline at end of file
+# mindtext.dataset.test1  
+
+Dataset的构建  
+=
+
+Example SST-2数据集Dataset构建
+
+    from mindtext.dataset.classification import SST2Dataset
+
+    dataset = SST2Dataset(paths='./mindtext/dataset/SST-2',
+                          tokenizer="./mindtext/pretrain/roberta-base",
+                          max_length=128,
+                          truncation_strategy=True,
+                          batch_size=32)
+    
+    ds = dataset() 
+    ds = dataset.from_cache( columns_list=['input_ids', 'attention_mask','label'],
+                          test_columns_list=['input_ids', 'attention_mask'],
+                          batch_size=32
+                          )
+    
+    train_dataset = ds['train']
+mindtext.dataset.base_dataset.Dataset
+--  
+>  class mindtext.dataset.base_dataset.Dataset( vocab (Vocabulary, Optional): Convert tokens to index,default None.
+>  name (str, Optional): Dataset name,default None.
+>  label_map (Dict[str, int], Optional): Dataset label map,default None.)  
+
+通过base_dataset中基类Dataset来构建文本分类、文本匹配和生成任务对应的数据集  
+
+>init(self, vocab: Vocabulary = None, name: str = None,
+                 label_map: Dict[str, int] = None)
+
+参数
+> + vocab(Vocabulary):  词表，默认为None  
+> + name(str):  下游任务数据集Dataset名称，默认为None  
+> + label_map(Dict[str, int], Optional):Dataset标签映射
+> 
+
+mindtext.dataset.base_dataset.CLSBaseDataset
+--
+
+文本分类Dataset的基类
+
+Example
+>class SST2Dataset(CLSBaseDataset):
+
+mindtext.dataset.base_dataset.PairCLSBaseDataset
+-
+
+文本匹配Dataset基类  
+
+Example  
+
+>class LCQMCDataset(PairCLSBaseDataset): 
\ No newline at end of file
diff --git a/docs/source/apis/mindtext.embeddings/index.rst b/docs/source/apis/mindtext.embeddings/index.rst
index 1af9f7c..87dc07d 100644
--- a/docs/source/apis/mindtext.embeddings/index.rst
+++ b/docs/source/apis/mindtext.embeddings/index.rst
@@ -9,11 +9,7 @@ mindtext.embeddings
 .. toctree::
    :maxdepth: 1
 
-   mindtext.embeddings.bert_embedding
    mindtext.embeddings.char_embedding
    mindtext.embeddings.embedding
-   mindtext.embeddings.luke_embedding
-   mindtext.embeddings.region_embedding
-   mindtext.embeddings.roberta_embedding
    mindtext.embeddings.static_embedding
 
diff --git a/docs/source/user/installation.rst b/docs/source/user/installation.rst
index 5a12c6e..663ffd0 100644
--- a/docs/source/user/installation.rst
+++ b/docs/source/user/installation.rst
@@ -19,10 +19,11 @@ mindtext 依赖如下包::
     numpy==1.21.2
     mindspore_gpu==1.3.0
 
-其中PyTorch的安装可能与操作系统及 CUDA 的版本相关，请参见 `PyTorch 官网 <https://pytorch.org/>`_ 。
-在依赖包安装完成的情况，您可以在命令行执行如下指令完成安装
+其中`mindspore`的安装可能与操作系统及 `CUDA` 的版本相关，请参见 `MindSpore 官网 <https://www.mindspore.cn/>`_ 。
+在依赖包安装完成的情况，您可以在命令行执行如下指令完成mindtext的安装
 
 ..  code:: shell
 
-   >>> pip install mindtext
-   >>> python -m spacy download en
+   >>> git clone https://gitee.com/mindspore/mindtext.git
+   >>> cd mindtext
+   >>> python setup.py install
-- 
Gitee