diff --git a/model_zoo/official/cv/centerface/README.md b/model_zoo/official/cv/centerface/README.md
index 82cb07b30bb..3eb77cca2af 100644
--- a/model_zoo/official/cv/centerface/README.md
+++ b/model_zoo/official/cv/centerface/README.md
@@ -48,7 +48,7 @@ Four loss is presented, total loss is their weighted mean.
 
 Note that you can run the scripts based on the dataset mentioned in original paper or widely used in relevant domain/network architecture. In the following sections, we will introduce how to run the scripts using the related dataset below.
 
-Dataset support: [WiderFace] or datasetd with the same format as WiderFace
+Dataset support: [WiderFace] or dataset with the same format as WiderFace
 Annotation support: [WiderFace] or annotation as the same format as WiderFace
 
 - The directory structure is as follows, the name of directory and file is user define:
@@ -105,7 +105,15 @@ step1: prepare pretrained model: train a mobilenet_v2 model by mindspore or use
 python convert_weight_mobilenetv2.py --ckpt_fn=./mobilenet_v2_key.ckpt --pt_fn=./mobilenet_v2-b0353104.pth --out_ckpt_fn=./mobilenet_v2.ckpt
 ```
 
-step2: prepare user rank_table
+step2: prepare dataset  
+
+ 1)download dataset from [WIDER FACE](http://mmlab.ie.cuhk.edu.hk/projects/WIDERFace/).  
+
+ 2)download the ground_truth from [ground_truth](http://mmlab.ie.cuhk.edu.hk/projects/WIDERFace/support/eval_script/eval_tools.zip).  
+
+ 3)download training annotations from [annotations](https://pan.baidu.com/s/1j_2wggZ3bvCuOAfZvjWqTg).  password: **f9hh**
+
+step3: prepare user rank_table
 
 ```python
 # user can use your own rank table file
@@ -114,7 +122,7 @@ step2: prepare user rank_table
 python hccl_tools.py --device_num "[0,8)"
 ```
 
-step3: train
+step4: train
 
 ```python
 cd scripts;
@@ -134,7 +142,7 @@ mkdir ./model
 cp device0/outputs/*/*.ckpt ./model # cp model to [MODEL_PATH]
 ```
 
-step4: test
+step5: test
 
 ```python
 # test CenterFace preparing
@@ -157,7 +165,7 @@ ls ./dataset/centerface/ground_truth/val.mat # annot_path
 sh test_distribute.sh
 ```
 
-step5: eval
+step6: eval
 
 ```python
 # after test, eval CenterFace, get MAP
diff --git a/model_zoo/official/cv/ssd/README_CN.md b/model_zoo/official/cv/ssd/README_CN.md
index 422a8274897..ed4d573e92e 100644
--- a/model_zoo/official/cv/ssd/README_CN.md
+++ b/model_zoo/official/cv/ssd/README_CN.md
@@ -42,10 +42,10 @@ SSD方法基于前向卷积网络，该网络产生固定大小的边界框集
 
 我们提供了4种不同的基础架构:
 
-- **ssd300**, 参考论文实现. 使用mobilenet-v2作为骨干网络, 并使用相同的bbox预测器.
-- ***ssd-mobilenet-v1-fpn**, 使用mobilenet-v1和FPN作为特征提取器, 并使用权重共享bbox预测器.
-- ***ssd-resnet50-fpn**, 使用resnet50和FPN作为特征提取器, 并使用权重共享bbox预测器.
-- **ssd-vgg16**, 参考论文实现. 使用vgg16作为骨干网络, 并使用相同的bbox预测器.
+- **ssd300**， 参考论文实现。 使用mobilenet-v2作为骨干网络， 并使用和论文相同的bbox预测器。
+- ***ssd-mobilenet-v1-fpn**， 使用mobilenet-v1和FPN作为特征提取器， 并使用权重共享box预测器。
+- ***ssd-resnet50-fpn**， 使用resnet50和FPN作为特征提取器， 并使用权重共享box预测器。
+- **ssd-vgg16**， 参考论文实现。 使用vgg16作为骨干网络， 并使用和论文相同的bbox预测器。
 
 # 数据集