update readme

Author: JasonWLM

.DS_Store

@@ -433,6 +433,16 @@ d. 清洗类-特殊逻辑规则：此类规则用于清洗一些特殊现象数
 
 ## 测评
 
+V2版SFT和base模型测评结果
+
+![image](image/evaluation_sft_v2.jpg)
+
+![image](image/evaluation_base_v2.jpg)
+
+<details>
+
+V1版SFT和base模型测评结果
+
 在 7 项英文 NLP 任务上，对 SFT 模型进行测评，以 OpenAI-InstructGPT-6B-SFT 为基准，归一化并平均各模型的得分，结果如下：
 
 ![image](image/auto-valuation-1.png)
@@ -441,6 +451,8 @@ d. 清洗类-特殊逻辑规则：此类规则用于清洗一些特殊现象数
 
 ![image](image/auto-valuation-2.png)
 
+</details>
+
 ## API
 
 TigerBot 提供了三种 API，包括 Chat-API，Plug-ins，Fine-Tunes。

README.md

@@ -338,6 +338,16 @@ We clean and filter data as follows:
   | [Wiki](https://huggingface.co/datasets/TigerResearch/tigerbot-wiki-plugin)                   | 100K     |
 
 ## Evaluation
+
+Evaluation result of V2 version SFT and and base model
+
+![image](image/evaluation_sft_v2.jpg)
+
+![image](image/evaluation_base_v2.jpg)
+
+<details>
+Evaluation result of V1 version SFT and and base model
+
 We evaluate our SFT models on seven public NLP datasets, and compare these with OpenAI-InstructGPT. 
 Results against OpenAI-InstructGPT-6B-SFT.
 
@@ -348,6 +358,8 @@ Results against bloom-7b1.
 
 ![image](image/auto-valuation-2.png)
 
+</details>
+
 ## API
 
 TigerBot provide APIs including Chat-API，Plug-ins，Fine-Tunes.