diff --git a/README.md b/README.md
index 9b42721..c381d2e 100644
--- a/README.md
+++ b/README.md
@@ -45,13 +45,6 @@
   </a>
 </div>
 
-## Table of Contents
-1. [Introduction](#1-introduction)
-2. [Release](#2-release)
-3. [Model Download](#3-model-download)
-4. [Quick Start](#4-quick-start)
-5. [License](#5-license)
-6. [Citation](#6-citation)
 
 <p align="center">
   <a href="https://github.com/deepseek-ai/DeepSeek-VL2/tree/main?tab=readme-ov-file#3-model-download"><b>📥 Model Download</b></a> |
@@ -63,6 +56,16 @@
   <a href=""><b>👁️ Demo</b></a>
 </p>
 
+## Table of Contents
+1. [Introduction](#1-introduction)
+2. [Release](#2-release)
+3. [Model Download](#3-model-download)
+4. [Quick Start](#4-quick-start)
+5. [License](#5-license)
+6. [Citation](#6-citation)
+
+
+
 ## 1. Introduction
 
 Introducing DeepSeek-VL2, an advanced series of large Mixture-of-Experts (MoE) Vision-Language Models that significantly improves upon its predecessor, DeepSeek-VL. DeepSeek-VL2 demonstrates superior capabilities across various tasks, including but not limited to visual question answering, optical character recognition,  document/table/chart understanding, and visual grounding. Our model series is composed of three variants: DeepSeek-VL2-Tiny, DeepSeek-VL2-Small and DeepSeek-VL2, with 1.0B, 2.8B and 4.5B activated parameters respectively.