diff --git "a/docs/C3/1. \350\207\252\345\256\232\344\271\211\345\257\274\345\205\245\346\250\241\345\236\213.md" "b/docs/C3/1. \350\207\252\345\256\232\344\271\211\345\257\274\345\205\245\346\250\241\345\236\213.md"
index ad2d1af..a15f027 100644
--- "a/docs/C3/1. \350\207\252\345\256\232\344\271\211\345\257\274\345\205\245\346\250\241\345\236\213.md"	
+++ "b/docs/C3/1. \350\207\252\345\256\232\344\271\211\345\257\274\345\205\245\346\250\241\345\236\213.md"	
@@ -113,7 +113,7 @@ llama.cpp是GGUF的开源项目，提供CLI和Server功能。
 ### 3.1 从HuggingFace下载Model
 
 最直觉的下载方式是通过git clone或者链接来下载，但是因为llm每部分都按GB计算，避免出现`OOM Error(Out of memory)`，我们可以使用Python写一个简单的download.py
-首先应该去hf拿到用户个人的`ACCESS_TOKEN`，打开 huggingface 。
+首先应该去hf拿到用户个人的`ACCESS_TOKEN`，打开 huggingface个人设置页面。
 
 ![alt text](../images/C3-3-1.png)
 
diff --git a/docs/images/C3-3-1.png b/docs/images/C3-3-1.png
index b1958db..a218e33 100644
Binary files a/docs/images/C3-3-1.png and b/docs/images/C3-3-1.png differ
diff --git a/docs/images/C3-3-2.png b/docs/images/C3-3-2.png
index 2c4a9fb..b575284 100644
Binary files a/docs/images/C3-3-2.png and b/docs/images/C3-3-2.png differ
diff --git a/docs/images/C3-3-3.png b/docs/images/C3-3-3.png
index 8ac8053..99c1cec 100644
Binary files a/docs/images/C3-3-3.png and b/docs/images/C3-3-3.png differ
diff --git "a/notebook/C3/1.\344\273\216GGUF\347\233\264\346\216\245\345\257\274\345\205\245/main.ipynb" "b/notebook/C3/1.\344\273\216GGUF\347\233\264\346\216\245\345\257\274\345\205\245/main.ipynb"
index 1cc3db2..9b5ff61 100644
--- "a/notebook/C3/1.\344\273\216GGUF\347\233\264\346\216\245\345\257\274\345\205\245/main.ipynb"
+++ "b/notebook/C3/1.\344\273\216GGUF\347\233\264\346\216\245\345\257\274\345\205\245/main.ipynb"
@@ -2,29 +2,9 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stderr",
-     "output_type": "stream",
-     "text": [
-      "\u001b[?25ltransferring model data ⠙ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠙ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠸ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠸ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠴ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠴ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠧ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠧ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠇ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠏ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠙ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠙ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠸ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠸ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠴ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠴ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠦ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠇ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠇ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠋ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠋ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠹ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠹ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data ⠼ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1Gtransferring model data \n",
-      "using existing layer sha256:99f013fc74fcfaab19a9cb36de4ebb8c50e9a15048ef88da2387ee7a0c0cffcb \n",
-      "using autodetected template chatml \n",
-      "using existing layer sha256:f02dd72bb2423204352eabc5637b44d79d17f109fdb510a7c51455892aa2d216 \n",
-      "creating new layer sha256:fb05d8484292a164d767deb5cd552e96b9fb4968bbcb855b0ae43cf7beb4e516 \n",
-      "writing manifest \n",
-      "success ⠋ \u001b[?25h\u001b[?25l\u001b[2K\u001b[1G\u001b[A\u001b[2K\u001b[1G\u001b[A\u001b[2K\u001b[1G\u001b[A\u001b[2K\u001b[1G\u001b[A\u001b[2K\u001b[1G\u001b[A\u001b[2K\u001b[1G\u001b[A\u001b[2K\u001b[1Gtransferring model data \n",
-      "using existing layer sha256:99f013fc74fcfaab19a9cb36de4ebb8c50e9a15048ef88da2387ee7a0c0cffcb \n",
-      "using autodetected template chatml \n",
-      "using existing layer sha256:f02dd72bb2423204352eabc5637b44d79d17f109fdb510a7c51455892aa2d216 \n",
-      "creating new layer sha256:fb05d8484292a164d767deb5cd552e96b9fb4968bbcb855b0ae43cf7beb4e516 \n",
-      "writing manifest \n",
-      "success \u001b[?25h\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# 1. 创建模型\n",
     "!ollama create mymodel -f Modelfile"
@@ -32,20 +12,9 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "NAME           \tID          \tSIZE  \tMODIFIED       \n",
-      "mymodel:latest \tb82a6e9999d2\t355 MB\t15 seconds ago\t\n",
-      "llama3:latest  \t365c0bd3c000\t4.7 GB\t13 hours ago  \t\n",
-      "llama3.1:latest\t75382d0899df\t4.7 GB\t15 hours ago  \t\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# 2.查看模型\n",
     "!ollama list"
@@ -53,20 +22,12 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
+   "execution_count": null,
    "metadata": {},
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "^C\n"
-     ]
-    }
-   ],
+   "outputs": [],
    "source": [
     "# 3. 终端内运行下列脚本运行模型\n",
-    "# ollama run mymodel"
+    "！ollama run mymodel"
    ]
   }
  ],