diff --git a/README.md b/README.md
index 86df35c..c66ace7 100755
--- a/README.md
+++ b/README.md
@@ -59,7 +59,7 @@ build_windows.bat
 
 Once the compilation is sucessfully finished, look for the executable file KSAI_Toolkits_demo in /output/$platform$/bin/ folder. You can execute it with the following command line paremeters:
 ```
-./output/linux/bin/KSAI_Toolkits_demo data/OCR/model/db_171_varied.tflite data/OCR/model/cls_best.tflite data/OCR/model/mv3_345000.tflite data/OCR/dictionary/char_6599.txt data/OCR/input/test.jpg
+./output/linux/bin/KSAI_Toolkits_demo data/OCR/model/ks_ocr_det_infer_v1.tflite data/OCR/model/ks_ocr_cls_infer_v1.tflite data/OCR/model/ks_ocr_recog_infer_v1.tflite data/OCR/dictionary/char_6599.txt data/OCR/input/test.jpg
 ```
 The output of the program should be:
 ```
diff --git a/doc/benchmark.md b/doc/benchmark.md
index 297c612..0d4ef21 100755
--- a/doc/benchmark.md
+++ b/doc/benchmark.md
@@ -106,4 +106,9 @@ xnnpack是google推出的针对CPU实现高度优化的浮点运算，针对移
 ### 总体性能统计
 ## OCR数据集
 * 图像数据(IP : https://sdk.ai.wpscdn.cn/KSAI/OCR_data/input/image.zip)
-* label信息(IP : https://sdk.ai.wpscdn.cn/KSAI/OCR_data/input/golden/annotation_jsonfile.zip)
\ No newline at end of file
+* label信息(IP : https://sdk.ai.wpscdn.cn/KSAI/OCR_data/input/golden/annotation_jsonfile.zip)
+## OCR性能统计
+|  OCR版本  |  检测（h1）  |  识别（带标点）  |  识别（不带标点）  |  总体recall（samples）  | 总体recall（words）  |
+| :------------: | :------------: | :------------: | :------------: | :------------: | :------------: |
+|  V1  |  0.788  |  0.89849  |  0.94123  |  0.85654  |  0.85078  |
+|  V2  |  0.788  |  0.91953  |  0.96091  |  0.8804  |  0.87982  |
\ No newline at end of file