{
  "version": "https://jsonfeed.org/version/1", 
  "title": "TensorFlow", 
  "description": "An open-source software library for Machine Intelligence", 
  "home_page_url": "https://www.v2ex.com/go/tensorflow", 
  "feed_url": "https://www.v2ex.com/feed/tensorflow.json", 
  "icon": "https://cdn.v2ex.com/navatar/4311/359e/974_large.png?m=1620107198", 
  "favicon": "https://cdn.v2ex.com/navatar/4311/359e/974_normal.png?m=1620107198", 
  "items": [
    {
      "author": {
        "url": "https://www.v2ex.com/member/morencato1", 
        "name": "morencato1", 
        "avatar": "https://cdn.v2ex.com/gravatar/ba9f6bd13ec4d3ab23e8b3431a310484?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/1144352", 
      "title": "\u5bb9\u5668\u5316\u90e8\u7f72\u6ca1\u6709\u9650\u5236\u4f7f\u7528 CPU \u548c\u5185\u5b58\u5927\u5c0f ollama \u53ea\u80fd\u8c03\u52a8\u4e00\u534a CPU \u600e\u4e48\u89e3\u51b3", 
      "id": "https://www.v2ex.com/t/1144352", 
      "date_published": "2025-07-10T09:53:55+00:00", 
      "content_html": "<h1>\u5bb9\u5668\u5316\u90e8\u7f72\u6ca1\u6709\u9650\u5236\u4f7f\u7528 CPU \u548c\u5185\u5b58\u5927\u5c0f ollama \u53ea\u80fd\u8c03\u52a8\u4e00\u534a CPU \u600e\u4e48\u89e3\u51b3</h1>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/morencato1", 
        "name": "morencato1", 
        "avatar": "https://cdn.v2ex.com/gravatar/ba9f6bd13ec4d3ab23e8b3431a310484?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/1144296", 
      "title": "\u300c\u63d0\u793a\u8bcd\u5fae\u8c03\u300d\u300cTIG-3.6 Mirage\u300dTIG-3.6-Mirage \u6280\u672f\u767d\u76ae\u4e66\uff1a\u63d0\u793a\u5fae\u8c03\u65b0\u8303\u5f0f\u7684\u5de5\u7a0b\u5b9e\u73b0\u4e0e\u4ea7\u4e1a\u4ef7\u503c", 
      "id": "https://www.v2ex.com/t/1144296", 
      "date_published": "2025-07-10T07:56:36+00:00", 
      "content_html": "<h1>TIG-3.6-Mirage \u6280\u672f\u767d\u76ae\u4e66\uff1a\u63d0\u793a\u5fae\u8c03\u65b0\u8303\u5f0f\u7684\u5de5\u7a0b\u5b9e\u73b0\u4e0e\u4ea7\u4e1a\u4ef7\u503c</h1>\n<p><strong>\u4f5c\u8005\uff1a</strong> \u5e7b\u5b99\u667a\u80fd\u56e2\u961f&amp;TIG-3.6-Mirage\n<strong>\u7248\u672c</strong>\uff1a2.0\n<strong>\u53d1\u5e03\u65f6\u95f4</strong>\uff1a2025 \u5e74 6 \u6708 8 \u65e5</p>\n<blockquote>\n<p>\u672c\u6587\u6863\u9488\u5bf9\u4e13\u4e1a\u793e\u533a\u8bb2\u89e3\u539f\u7406\u548c\u6280\u672f\u5b9e\u73b0\uff0c\u5982\u679c\u60a8\u662f\u666e\u901a\u7528\u6237\u8bf7<a href=\"https://bbs.furina.chat/t/topic/198\" rel=\"nofollow\">\u70b9\u6b64\u5feb\u901f\u4e0a\u624b</a></p>\n</blockquote>\n<p><a href=\"https://github.com/PhantasmAI/PromptMicroTune-Revolution\" rel=\"nofollow\">https://github.com/PhantasmAI/PromptMicroTune-Revolution</a></p>\n<hr/>\n<h2>\u6458\u8981</h2>\n<p>\u672c\u767d\u76ae\u4e66\u8be6\u7ec6\u9610\u8ff0\u4e86 TIG-3.6-Mirage \u6240\u91c7\u7528\u7684\u63d0\u793a\u5fae\u8c03\uff08 Prompt Fine-tuning \uff09\u6280\u672f\u8303\u5f0f\u7684\u5de5\u7a0b\u5b9e\u73b0\u7ec6\u8282\u4e0e\u4ea7\u4e1a\u4ef7\u503c\u3002\u8be5\u6280\u672f\u57fa\u4e8e few-shots \u5b66\u4e60\u548c Chain of Thought \uff08 CoT \uff09\u7684\u6df1\u5ea6\u878d\u5408\uff0c\u901a\u8fc7\"\u4e09\u6bb5\u5f0f\u8ba4\u77e5\u6a21\u677f\"\u5b9e\u73b0\u4e86\u8f6f\u5fae\u8c03\u7684\u5de5\u7a0b\u5316\u843d\u5730\u3002\u6211\u4eec\u5766\u8bda\u627f\u8ba4\u8fd9\u4e00\u6280\u672f\u5efa\u7acb\u5728\u4e1a\u754c\u6210\u719f\u7684 few-shots \u548c CoT \u57fa\u7840\u4e4b\u4e0a\uff0c\u4f46\u901a\u8fc7\u7cfb\u7edf\u6027\u7684\u5de5\u7a0b\u521b\u65b0\u548c\u8ba4\u77e5\u67b6\u6784\u8bbe\u8ba1\uff0c\u5b9e\u73b0\u4e86\"\u7528\u6237\u96f6\u95e8\u69db\u4f7f\u7528\uff0c\u7cfb\u7edf\u627f\u62c5\u5168\u90e8\u590d\u6742\u6027\"\u7684\u6280\u672f\u7406\u5ff5\u3002\u5728\u4fdd\u6301 99%\u4ee5\u4e0a\u8ba1\u7b97\u6210\u672c\u964d\u4f4e\u7684\u540c\u65f6\uff0c\u6211\u4eec\u6784\u5efa\u4e86\u5b8c\u6574\u7684\u8ba4\u77e5\u5f15\u64ce\u6765\u5904\u7406\u89d2\u8272\u4e00\u81f4\u6027\u3001\u60c5\u611f\u8fde\u8d2f\u6027\u3001\u957f\u7a0b\u8bb0\u5fc6\u7b49\u590d\u6742\u95ee\u9898\uff0c\u4e3a AI \u5b9a\u5236\u5316\u5e94\u7528\u5f00\u8f9f\u4e86\u5168\u65b0\u7684\u6280\u672f\u8def\u5f84\u3002</p>\n<h2>1. \u6280\u672f\u80cc\u666f\uff1a\u4ece\u53c2\u6570\u5fae\u8c03\u5230\u8ba4\u77e5\u5fae\u8c03\u7684\u8303\u5f0f\u9769\u547d</h2>\n<h3>1.1 \u4f20\u7edf\u5fae\u8c03\u65b9\u6cd5\u7684\u5de5\u7a0b\u74f6\u9888</h3>\n<p>\u5f53\u524d\u4e3b\u6d41\u7684\u6a21\u578b\u5b9a\u5236\u5316\u65b9\u6cd5\u4e3b\u8981\u4f9d\u8d56\u53c2\u6570\u7ea7\u5fae\u8c03\uff0c\u5305\u62ec\u5168\u53c2\u6570\u5fae\u8c03\u3001LoRA \u3001Adapter \u7b49\u6280\u672f\u8def\u7ebf\u3002\u8fd9\u4e9b\u65b9\u6cd5\u5728\u5de5\u7a0b\u5b9e\u8df5\u4e2d\u9762\u4e34\u4e09\u4e2a\u6838\u5fc3\u74f6\u9888\uff1a</p>\n<p><strong>\u8ba1\u7b97\u8d44\u6e90\u74f6\u9888</strong>\uff1a\u57fa\u4e8e\u4e2d\u56fd\u5e02\u573a\u7684\u5b9e\u9645\u6570\u636e\uff0c\u4ee5 7B \u53c2\u6570\u6a21\u578b\u4e3a\u4f8b\uff0c\u5168\u53c2\u6570\u5fae\u8c03\u9700\u8981\u81f3\u5c11 4\u00d7A100 GPU \uff0c\u8bad\u7ec3\u65f6\u95f4 8-24 \u5c0f\u65f6\uff0c\u5355\u6b21\u6210\u672c\u7ea6 2000-3000 \u5143\u4eba\u6c11\u5e01\u3002\u5373\u4f7f\u91c7\u7528 LoRA \u7b49\u53c2\u6570\u9ad8\u6548\u65b9\u6cd5\uff0c\u4ecd\u9700\u8981 1\u00d7A10 \u6216 T4 GPU \u548c 4-8 \u5c0f\u65f6\u8bad\u7ec3\u65f6\u95f4\uff0c\u6210\u672c\u7ea6 300-600 \u5143\u4eba\u6c11\u5e01\u3002\u8fd9\u79cd\u6210\u672c\u7ed3\u6784\u4f7f\u5f97\u5927\u591a\u6570\u5f00\u53d1\u8005\u548c\u4e2d\u5c0f\u4f01\u4e1a\u65e0\u6cd5\u627f\u62c5\u9891\u7e41\u7684\u6a21\u578b\u5b9a\u5236\u9700\u6c42\u3002</p>\n<p><strong>\u8fed\u4ee3\u6548\u7387\u74f6\u9888</strong>\uff1a\u4f20\u7edf\u5fae\u8c03\u4ece\u6570\u636e\u51c6\u5907\u3001\u8bad\u7ec3\u6267\u884c\u5230\u6548\u679c\u9a8c\u8bc1\u7684\u5b8c\u6574\u5468\u671f\u901a\u5e38\u9700\u8981 1-3 \u5929\u3002\u5728\u5feb\u901f\u8fed\u4ee3\u7684\u4ea7\u54c1\u5f00\u53d1\u73af\u5883\u4e2d\uff0c\u8fd9\u79cd\u65f6\u95f4\u6210\u672c\u4e25\u91cd\u5236\u7ea6\u4e86\u521b\u65b0\u901f\u5ea6\u3002\u66f4\u5173\u952e\u7684\u662f\uff0c\u6bcf\u6b21\u9700\u6c42\u53d8\u66f4\u90fd\u9700\u8981\u91cd\u65b0\u6267\u884c\u5b8c\u6574\u7684\u8bad\u7ec3\u6d41\u7a0b\uff0c\u65e0\u6cd5\u652f\u6301\u654f\u6377\u5f00\u53d1\u6a21\u5f0f\u3002</p>\n<p><strong>\u6280\u672f\u95e8\u69db\u74f6\u9888</strong>\uff1a\u53c2\u6570\u5fae\u8c03\u9700\u8981\u6df1\u5ea6\u5b66\u4e60\u4e13\u4e1a\u77e5\u8bc6\uff0c\u5305\u62ec\u6570\u636e\u9884\u5904\u7406\u3001\u8d85\u53c2\u6570\u8c03\u4f18\u3001\u68af\u5ea6\u76d1\u63a7\u3001\u6536\u655b\u5224\u65ad\u7b49\u590d\u6742\u6280\u672f\u73af\u8282\u3002\u8fd9\u79cd\u6280\u672f\u590d\u6742\u6027\u5c06\u5927\u91cf\u6f5c\u5728\u7528\u6237\u6392\u9664\u5728\u5916\uff0c\u9650\u5236\u4e86 AI \u6280\u672f\u7684\u666e\u53ca\u5e94\u7528\u3002</p>\n<h3>1.2 \u8ba4\u77e5\u5fae\u8c03\u7684\u6280\u672f\u7a81\u7834</h3>\n<p>\u6211\u4eec\u63d0\u51fa\u7684\u8ba4\u77e5\u5fae\u8c03\u8303\u5f0f\u4ece\u6839\u672c\u4e0a\u6539\u53d8\u4e86\u6a21\u578b\u5b9a\u5236\u5316\u7684\u6280\u672f\u8def\u5f84\u3002\u6838\u5fc3\u601d\u60f3\u662f\u5c06\u5b9a\u5236\u5316\u9700\u6c42\u4ece\u53c2\u6570\u7a7a\u95f4\u8f6c\u79fb\u5230\u8ba4\u77e5\u7a7a\u95f4\uff0c\u901a\u8fc7\u7cbe\u5fc3\u8bbe\u8ba1\u7684\u8ba4\u77e5\u6a21\u677f\u6765\u5f15\u5bfc\u6a21\u578b\u884c\u4e3a\uff0c\u800c\u975e\u4fee\u6539\u6a21\u578b\u6743\u91cd\u3002</p>\n<p>\u8fd9\u79cd\u65b9\u6cd5\u7684\u6280\u672f\u4f18\u52bf\u4f53\u73b0\u5728\u4e09\u4e2a\u5c42\u9762\uff1a</p>\n<p><strong>\u6210\u672c\u4f18\u52bf</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u5b8c\u5168\u907f\u514d\u4e86 GPU \u8bad\u7ec3\u6210\u672c\uff0c\u4ec5\u5728\u63a8\u7406\u9636\u6bb5\u4ea7\u751f\u8c03\u7528\u8d39\u7528\u3002\u57fa\u4e8e TIG-3.6-Mirage \u7684\u5b9e\u9645\u5b9a\u4ef7\uff08 Input: \u00a542/M Tokens, Output: \u00a5126/M Tokens \uff09\uff0c\u4e00\u4e2a\u5b8c\u6574\u7684\u89d2\u8272\u5b9a\u5236\u9879\u76ee\u603b\u6210\u672c\u7ea6 50-200 \u5143\u4eba\u6c11\u5e01\uff0c\u76f8\u6bd4\u4f20\u7edf LoRA \u5fae\u8c03\u7684 300-600 \u5143\u4eba\u6c11\u5e01\u964d\u4f4e 70-85%\uff0c\u76f8\u6bd4\u5168\u53c2\u6570\u5fae\u8c03\u7684 2000-3000 \u5143\u4eba\u6c11\u5e01\u964d\u4f4e 90-95%\u3002\u6839\u636e\u5b98\u65b9\u6d4b\u8bd5\uff0c\u9996\u8f6e system \u63d0\u793a\u22651185 Token \u662f\u6548\u679c\u4e0e\u6210\u672c\u6700\u5e73\u8861\u7684\u533a\u95f4\u3002</p>\n<p><strong>\u6548\u7387\u4f18\u52bf</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u7684\u914d\u7f6e\u65f6\u95f4\u4ece\u4f20\u7edf\u7684\u5c0f\u65f6\u7ea7\u964d\u4f4e\u5230\u5206\u949f\u7ea7\u3002\u4e00\u4e2a\u7ecf\u9a8c\u4e30\u5bcc\u7684\u5de5\u7a0b\u5e08\u53ef\u4ee5\u5728 10-30 \u5206\u949f\u5185\u5b8c\u6210\u4e00\u4e2a\u590d\u6742\u89d2\u8272\u7684\u8ba4\u77e5\u914d\u7f6e\uff0c\u800c\u4f20\u7edf\u5fae\u8c03\u9700\u8981 4-8 \u5c0f\u65f6\u7684\u8bad\u7ec3\u65f6\u95f4\u52a0\u4e0a\u6570\u5c0f\u65f6\u7684\u8c03\u8bd5\u65f6\u95f4\u3002</p>\n<p><strong>\u95e8\u69db\u4f18\u52bf</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u5c06\u6280\u672f\u590d\u6742\u6027\u5b8c\u5168\u5c01\u88c5\u5728\u7cfb\u7edf\u5185\u90e8\uff0c\u7528\u6237\u53ea\u9700\u8981\u901a\u8fc7\u81ea\u7136\u8bed\u8a00\u63cf\u8ff0\u9700\u6c42\u5373\u53ef\u5b8c\u6210\u5b9a\u5236\u5316\u3002\u8fd9\u79cd\"\u5199\u4f5c\u5373\u7f16\u7a0b\"\u7684\u4ea4\u4e92\u6a21\u5f0f\u5c06 AI \u5b9a\u5236\u5316\u7684\u95e8\u69db\u964d\u4f4e\u5230\u666e\u901a\u5185\u5bb9\u521b\u4f5c\u8005\u7684\u6c34\u5e73\u3002</p>\n<h2>2. \u6838\u5fc3\u6280\u672f\u67b6\u6784\uff1a\u4e09\u6bb5\u5f0f\u8ba4\u77e5\u5f15\u64ce\u7684\u5de5\u7a0b\u5b9e\u73b0</h2>\n<h3>2.1 \u4e09\u6bb5\u5f0f\u8ba4\u77e5\u6a21\u677f\u7684\u8bbe\u8ba1\u539f\u7406</h3>\n<p>\u6211\u4eec\u7684\u6838\u5fc3\u6280\u672f\u521b\u65b0\u662f\"\u4e09\u6bb5\u5f0f\u8ba4\u77e5\u6a21\u677f\"\uff08 Three-Stage Cognitive Template \uff09\uff0c\u8fd9\u662f\u4e00\u79cd\u5c06 few-shots \u5b66\u4e60\u548c CoT \u63a8\u7406\u6df1\u5ea6\u878d\u5408\u7684\u8ba4\u77e5\u67b6\u6784\u3002\u8be5\u67b6\u6784\u5305\u542b\u4e09\u4e2a\u5173\u952e\u7ec4\u4ef6\uff1a</p>\n<p><strong>\u60c5\u5883\u5b9a\u4e49\u6bb5\uff08 Context Definition \uff09</strong>\uff1a\u660e\u786e\u5b9a\u4e49\u89d2\u8272\u7684\u57fa\u672c\u5c5e\u6027\u3001\u6240\u5904\u73af\u5883\u3001\u5f53\u524d\u72b6\u6001\u7b49\u5173\u952e\u4fe1\u606f\u3002\u8fd9\u4e00\u6bb5\u7684\u4f5c\u7528\u662f\u4e3a\u6a21\u578b\u5efa\u7acb\u6e05\u6670\u7684\u8ba4\u77e5\u951a\u70b9\uff0c\u786e\u4fdd\u540e\u7eed\u63a8\u7406\u8fc7\u7a0b\u6709\u660e\u786e\u7684\u53c2\u7167\u6846\u67b6\u3002</p>\n<p><strong>\u601d\u7ef4\u8fc7\u7a0b\u6bb5\uff08 Thinking Process \uff09</strong>\uff1a\u8fd9\u662f\u6574\u4e2a\u67b6\u6784\u7684\u6838\u5fc3\u521b\u65b0\u3002\u6211\u4eec\u5c06\u4f20\u7edf CoT \u7684\u903b\u8f91\u63a8\u7406\u6269\u5c55\u4e3a\u591a\u7ef4\u5ea6\u7684\u8ba4\u77e5\u6d41\u52a8\uff0c\u5305\u62ec\u60c5\u611f\u6ce2\u52a8\u3001\u8bb0\u5fc6\u8054\u60f3\u3001\u52a8\u673a\u5206\u6790\u3001\u51b2\u7a81\u89e3\u51b3\u7b49\u4eba\u7c7b\u601d\u7ef4\u7684\u590d\u6742\u7279\u5f81\u3002\u8fd9\u4e00\u6bb5\u901a\u8fc7\u81ea\u7136\u8bed\u8a00\u7684\u610f\u8bc6\u6d41\u5f62\u5f0f\uff0c\u8ba9\u6a21\u578b\u5b66\u4f1a\"\u50cf\u4eba\u4e00\u6837\u601d\u8003\"\u3002</p>\n<p><strong>\u884c\u4e3a\u8f93\u51fa\u6bb5\uff08 Behavioral Output \uff09</strong>\uff1a\u57fa\u4e8e\u524d\u4e24\u6bb5\u7684\u8ba4\u77e5\u57fa\u7840\uff0c\u751f\u6210\u7b26\u5408\u89d2\u8272\u7279\u5f81\u7684\u5177\u4f53\u884c\u4e3a\u548c\u8bed\u8a00\u8f93\u51fa\u3002\u8fd9\u4e00\u6bb5\u4e0d\u4ec5\u5305\u542b\u8868\u9762\u7684\u8bed\u8a00\u8868\u8fbe\uff0c\u8fd8\u5305\u542b\u52a8\u4f5c\u63cf\u8ff0\u3001\u60c5\u611f\u8868\u8fbe\u3001\u73af\u5883\u4e92\u52a8\u7b49\u591a\u6a21\u6001\u4fe1\u606f\u3002</p>\n<h3>2.2 \u8ba4\u77e5\u4e00\u81f4\u6027\u4fdd\u969c\u673a\u5236</h3>\n<p>\u5728\u4e09\u6bb5\u5f0f\u67b6\u6784\u57fa\u7840\u4e0a\uff0c\u6211\u4eec\u5f00\u53d1\u4e86\u4e00\u5957\u590d\u6742\u7684\u8ba4\u77e5\u4e00\u81f4\u6027\u4fdd\u969c\u673a\u5236\uff0c\u786e\u4fdd\u6a21\u578b\u5728\u957f\u671f\u4ea4\u4e92\u4e2d\u7ef4\u6301\u7a33\u5b9a\u7684\u89d2\u8272\u7279\u5f81\u3002</p>\n<p><strong>\u8bb0\u5fc6\u8fde\u8d2f\u6027\u7ba1\u7406</strong>\uff1a\u6211\u4eec\u8bbe\u8ba1\u4e86\u5206\u5c42\u8bb0\u5fc6\u67b6\u6784\uff0c\u5305\u62ec\u6838\u5fc3\u8bb0\u5fc6\uff08\u89d2\u8272\u7684\u57fa\u672c\u5c5e\u6027\u548c\u4ef7\u503c\u89c2\uff09\u3001\u60c5\u8282\u8bb0\u5fc6\uff08\u91cd\u8981\u7684\u4ea4\u4e92\u5386\u53f2\uff09\u3001\u5de5\u4f5c\u8bb0\u5fc6\uff08\u5f53\u524d\u5bf9\u8bdd\u7684\u4e0a\u4e0b\u6587\u4fe1\u606f\uff09\u3002\u7cfb\u7edf\u901a\u8fc7\u6ce8\u610f\u529b\u6743\u91cd\u5206\u914d\u673a\u5236\uff0c\u786e\u4fdd\u4e0d\u540c\u5c42\u6b21\u7684\u8bb0\u5fc6\u4fe1\u606f\u5f97\u5230\u5408\u9002\u7684\u6fc0\u6d3b\u548c\u6291\u5236\u3002</p>\n<p><strong>\u60c5\u611f\u72b6\u6001\u8ffd\u8e2a</strong>\uff1a\u6211\u4eec\u6784\u5efa\u4e86\u591a\u7ef4\u5ea6\u7684\u60c5\u611f\u72b6\u6001\u7a7a\u95f4\uff0c\u5305\u62ec\u57fa\u7840\u60c5\u611f\uff08\u559c\u6012\u54c0\u4e50\uff09\u3001\u590d\u5408\u60c5\u611f\uff08\u5ac9\u5992\u3001\u6000\u5ff5\u3001\u671f\u5f85\uff09\u3001\u60c5\u611f\u5f3a\u5ea6\u3001\u60c5\u611f\u53d8\u5316\u8d8b\u52bf\u7b49\u3002\u7cfb\u7edf\u901a\u8fc7\u60c5\u611f\u72b6\u6001\u5411\u91cf\u7684\u52a8\u6001\u66f4\u65b0\uff0c\u786e\u4fdd\u89d2\u8272\u7684\u60c5\u611f\u8868\u8fbe\u5177\u6709\u8fde\u8d2f\u6027\u548c\u771f\u5b9e\u6027\u3002</p>\n<p><strong>\u884c\u4e3a\u6a21\u5f0f\u7ea6\u675f</strong>\uff1a\u6211\u4eec\u5efa\u7acb\u4e86\u884c\u4e3a\u4e00\u81f4\u6027\u68c0\u67e5\u673a\u5236\uff0c\u901a\u8fc7\u9884\u5b9a\u4e49\u7684\u884c\u4e3a\u6a21\u5f0f\u5e93\u548c\u5b9e\u65f6\u884c\u4e3a\u5206\u6790\uff0c\u786e\u4fdd\u89d2\u8272\u7684\u884c\u4e3a\u9009\u62e9\u7b26\u5408\u5176\u4eba\u683c\u7279\u5f81\u3002\u5f53\u68c0\u6d4b\u5230\u884c\u4e3a\u504f\u79bb\u65f6\uff0c\u7cfb\u7edf\u4f1a\u81ea\u52a8\u89e6\u53d1\u4fee\u6b63\u673a\u5236\u3002</p>\n<h3>2.3 \u52a8\u6001\u9002\u5e94\u6027\u8c03\u8282\u7cfb\u7edf</h3>\n<p>\u4e3a\u4e86\u5904\u7406\u590d\u6742\u591a\u53d8\u7684\u4ea4\u4e92\u573a\u666f\uff0c\u6211\u4eec\u5f00\u53d1\u4e86\u52a8\u6001\u9002\u5e94\u6027\u8c03\u8282\u7cfb\u7edf\uff0c\u8be5\u7cfb\u7edf\u80fd\u591f\u6839\u636e\u7528\u6237\u53cd\u9988\u548c\u60c5\u5883\u53d8\u5316\u5b9e\u65f6\u8c03\u6574\u8ba4\u77e5\u7b56\u7565\u3002</p>\n<p><strong>\u7528\u6237\u753b\u50cf\u5206\u6790</strong>\uff1a\u7cfb\u7edf\u901a\u8fc7\u5bf9\u8bdd\u6a21\u5f0f\u5206\u6790\u3001\u53cd\u9988\u9891\u7387\u7edf\u8ba1\u3001\u4ea4\u4e92\u6df1\u5ea6\u8bc4\u4f30\u7b49\u591a\u4e2a\u7ef4\u5ea6\uff0c\u6784\u5efa\u7528\u6237\u7684\u7ecf\u9a8c\u753b\u50cf\u548c\u504f\u597d\u6a21\u578b\u3002\u57fa\u4e8e\u8fd9\u4e9b\u4fe1\u606f\uff0c\u7cfb\u7edf\u53ef\u4ee5\u52a8\u6001\u8c03\u6574\u89d2\u8272\u7684\u8868\u8fbe\u98ce\u683c\u3001\u4e92\u52a8\u6df1\u5ea6\u3001\u4e3b\u52a8\u6027\u7a0b\u5ea6\u7b49\u7279\u5f81\u3002</p>\n<p><strong>\u60c5\u5883\u611f\u77e5\u673a\u5236</strong>\uff1a\u7cfb\u7edf\u5177\u5907\u5bf9\u5bf9\u8bdd\u60c5\u5883\u7684\u5b9e\u65f6\u611f\u77e5\u80fd\u529b\uff0c\u5305\u62ec\u8bdd\u9898\u8f6c\u6362\u3001\u60c5\u611f\u6c1b\u56f4\u53d8\u5316\u3001\u7d27\u5f20\u7a0b\u5ea6\u5347\u7ea7\u7b49\u3002\u57fa\u4e8e\u60c5\u5883\u611f\u77e5\u7ed3\u679c\uff0c\u7cfb\u7edf\u4f1a\u76f8\u5e94\u8c03\u6574\u89d2\u8272\u7684\u8ba4\u77e5\u7b56\u7565\u548c\u884c\u4e3a\u6a21\u5f0f\u3002</p>\n<p><strong>\u81ea\u9002\u5e94\u5b66\u4e60\u673a\u5236</strong>\uff1a\u7cfb\u7edf\u901a\u8fc7\u5f3a\u5316\u5b66\u4e60\u7b97\u6cd5\uff0c\u4ece\u7528\u6237\u7684\u6b63\u9762\u548c\u8d1f\u9762\u53cd\u9988\u4e2d\u5b66\u4e60\u4f18\u5316\u7b56\u7565\u3002\u8fd9\u79cd\u5b66\u4e60\u4e0d\u4f1a\u4fee\u6539\u57fa\u7840\u6a21\u578b\u53c2\u6570\uff0c\u800c\u662f\u8c03\u6574\u8ba4\u77e5\u6a21\u677f\u7684\u6743\u91cd\u5206\u914d\u548c\u6fc0\u6d3b\u6a21\u5f0f\u3002</p>\n<h2>3. \u5de5\u7a0b\u5b9e\u73b0\u7ec6\u8282\uff1a\u4ece\u7406\u8bba\u5230\u751f\u4ea7\u7684\u6280\u672f\u6311\u6218</h2>\n<h3>3.1 \u63d0\u793a\u7a33\u5b9a\u6027\u5de5\u7a0b</h3>\n<p>\u5728\u751f\u4ea7\u73af\u5883\u4e2d\uff0c\u63d0\u793a\u7684\u5fae\u5c0f\u53d8\u5316\u53ef\u80fd\u5bfc\u81f4\u6a21\u578b\u8f93\u51fa\u7684\u663e\u8457\u6ce2\u52a8\uff0c\u8fd9\u662f\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u9762\u4e34\u7684\u6838\u5fc3\u5de5\u7a0b\u6311\u6218\u3002\u6211\u4eec\u901a\u8fc7\u591a\u5c42\u6b21\u7684\u7a33\u5b9a\u6027\u4fdd\u969c\u673a\u5236\u6765\u89e3\u51b3\u8fd9\u4e2a\u95ee\u9898\u3002</p>\n<p><strong>\u6a21\u677f\u6807\u51c6\u5316\u4f53\u7cfb</strong>\uff1a\u6211\u4eec\u5efa\u7acb\u4e86\u4e25\u683c\u7684\u8ba4\u77e5\u6a21\u677f\u8bbe\u8ba1\u89c4\u8303\uff0c\u5305\u62ec\u8bed\u8a00\u98ce\u683c\u6307\u5357\u3001\u7ed3\u6784\u5316\u8981\u6c42\u3001\u5173\u952e\u8bcd\u4f7f\u7528\u89c4\u5219\u7b49\u3002\u6bcf\u4e2a\u6a21\u677f\u90fd\u9700\u8981\u901a\u8fc7\u6807\u51c6\u5316\u68c0\u67e5\uff0c\u786e\u4fdd\u5176\u7b26\u5408\u7cfb\u7edf\u7684\u8ba4\u77e5\u67b6\u6784\u8981\u6c42\u3002</p>\n<p><strong>\u9c81\u68d2\u6027\u6d4b\u8bd5\u6846\u67b6</strong>\uff1a\u6211\u4eec\u5f00\u53d1\u4e86\u81ea\u52a8\u5316\u7684\u9c81\u68d2\u6027\u6d4b\u8bd5\u5de5\u5177\uff0c\u80fd\u591f\u5bf9\u8ba4\u77e5\u6a21\u677f\u8fdb\u884c\u5927\u89c4\u6a21\u7684\u53d8\u5f02\u6d4b\u8bd5\u3002\u8be5\u5de5\u5177\u901a\u8fc7\u540c\u4e49\u8bcd\u66ff\u6362\u3001\u53e5\u5f0f\u53d8\u6362\u3001\u987a\u5e8f\u8c03\u6574\u7b49\u65b9\u5f0f\u751f\u6210\u6d4b\u8bd5\u7528\u4f8b\uff0c\u8bc4\u4f30\u6a21\u677f\u7684\u7a33\u5b9a\u6027\u8868\u73b0\u3002</p>\n<p><strong>\u591a\u7248\u672c\u96c6\u6210\u7b56\u7565</strong>\uff1a\u4e3a\u4e86\u8fdb\u4e00\u6b65\u63d0\u5347\u7a33\u5b9a\u6027\uff0c\u6211\u4eec\u91c7\u7528\u4e86\u591a\u7248\u672c\u96c6\u6210\u7684\u6280\u672f\u65b9\u6848\u3002\u7cfb\u7edf\u4f1a\u4e3a\u6bcf\u4e2a\u89d2\u8272\u7ef4\u62a4\u591a\u4e2a\u8bed\u4e49\u7b49\u4ef7\u4f46\u8868\u8fbe\u4e0d\u540c\u7684\u8ba4\u77e5\u6a21\u677f\uff0c\u901a\u8fc7\u96c6\u6210\u5b66\u4e60\u7684\u65b9\u5f0f\u4ea7\u751f\u6700\u7ec8\u8f93\u51fa\u3002</p>\n<h3>3.2 \u6027\u80fd\u4f18\u5316\u4e0e\u6269\u5c55\u6027\u8bbe\u8ba1</h3>\n<p>\u8ba4\u77e5\u5fae\u8c03\u867d\u7136\u907f\u514d\u4e86\u8bad\u7ec3\u6210\u672c\uff0c\u4f46\u5728\u63a8\u7406\u9636\u6bb5\u4f1a\u589e\u52a0\u4e00\u5b9a\u7684\u8ba1\u7b97\u5f00\u9500\u3002\u6211\u4eec\u901a\u8fc7\u591a\u79cd\u6280\u672f\u624b\u6bb5\u6765\u4f18\u5316\u7cfb\u7edf\u6027\u80fd\u3002</p>\n<p><strong>\u6a21\u677f\u7f13\u5b58\u673a\u5236</strong>\uff1a\u6211\u4eec\u8bbe\u8ba1\u4e86\u667a\u80fd\u7684\u6a21\u677f\u7f13\u5b58\u7cfb\u7edf\uff0c\u5bf9\u9891\u7e41\u4f7f\u7528\u7684\u8ba4\u77e5\u6a21\u677f\u8fdb\u884c\u9884\u5904\u7406\u548c\u7f13\u5b58\u3002\u8fd9\u79cd\u673a\u5236\u53ef\u4ee5\u5c06\u6a21\u677f\u52a0\u8f7d\u65f6\u95f4\u4ece\u6beb\u79d2\u7ea7\u964d\u4f4e\u5230\u5fae\u79d2\u7ea7\uff0c\u663e\u8457\u63d0\u5347\u54cd\u5e94\u901f\u5ea6\u3002</p>\n<p><strong>\u5e76\u884c\u5904\u7406\u67b6\u6784</strong>\uff1a\u6211\u4eec\u91c7\u7528\u4e86\u5f02\u6b65\u5e76\u884c\u7684\u5904\u7406\u67b6\u6784\uff0c\u5c06\u8ba4\u77e5\u6a21\u677f\u7684\u89e3\u6790\u3001\u7528\u6237\u8f93\u5165\u7684\u5206\u6790\u3001\u8f93\u51fa\u751f\u6210\u7b49\u73af\u8282\u5e76\u884c\u5316\u5904\u7406\u3002\u8fd9\u79cd\u67b6\u6784\u8bbe\u8ba1\u4f7f\u5f97\u7cfb\u7edf\u80fd\u591f\u5728\u4fdd\u6301\u9ad8\u8d28\u91cf\u8f93\u51fa\u7684\u540c\u65f6\uff0c\u7ef4\u6301\u8f83\u5feb\u7684\u54cd\u5e94\u901f\u5ea6\u3002</p>\n<p><strong>\u5f39\u6027\u6269\u5c55\u8bbe\u8ba1</strong>\uff1a\u6211\u4eec\u7684\u7cfb\u7edf\u91c7\u7528\u4e86\u5fae\u670d\u52a1\u67b6\u6784\uff0c\u652f\u6301\u6839\u636e\u8d1f\u8f7d\u60c5\u51b5\u52a8\u6001\u6269\u5c55\u3002\u8ba4\u77e5\u5f15\u64ce\u3001\u6a21\u677f\u7ba1\u7406\u3001\u7528\u6237\u753b\u50cf\u5206\u6790\u7b49\u6a21\u5757\u90fd\u53ef\u4ee5\u72ec\u7acb\u6269\u5c55\uff0c\u786e\u4fdd\u7cfb\u7edf\u80fd\u591f\u5e94\u5bf9\u5927\u89c4\u6a21\u5e76\u53d1\u8bbf\u95ee\u3002</p>\n<h3>3.3 \u8d28\u91cf\u4fdd\u969c\u4e0e\u76d1\u63a7\u4f53\u7cfb</h3>\n<p>\u4e3a\u4e86\u786e\u4fdd\u751f\u4ea7\u73af\u5883\u4e2d\u7684\u8f93\u51fa\u8d28\u91cf\uff0c\u6211\u4eec\u5efa\u7acb\u4e86\u5b8c\u6574\u7684\u8d28\u91cf\u4fdd\u969c\u4e0e\u76d1\u63a7\u4f53\u7cfb\u3002</p>\n<p><strong>\u5b9e\u65f6\u8d28\u91cf\u76d1\u63a7</strong>\uff1a\u7cfb\u7edf\u90e8\u7f72\u4e86\u591a\u7ef4\u5ea6\u7684\u8d28\u91cf\u76d1\u63a7\u6307\u6807\uff0c\u5305\u62ec\u89d2\u8272\u4e00\u81f4\u6027\u5f97\u5206\u3001\u60c5\u611f\u9002\u914d\u5ea6\u3001\u8bed\u8a00\u6d41\u7545\u5ea6\u3001\u7528\u6237\u6ee1\u610f\u5ea6\u7b49\u3002\u5f53\u4efb\u4f55\u6307\u6807\u51fa\u73b0\u5f02\u5e38\u65f6\uff0c\u7cfb\u7edf\u4f1a\u81ea\u52a8\u89e6\u53d1\u544a\u8b66\u548c\u4fee\u6b63\u673a\u5236\u3002</p>\n<p><strong>A/B \u6d4b\u8bd5\u6846\u67b6</strong>\uff1a\u6211\u4eec\u5efa\u7acb\u4e86\u5b8c\u6574\u7684 A/B \u6d4b\u8bd5\u6846\u67b6\uff0c\u652f\u6301\u5bf9\u4e0d\u540c\u8ba4\u77e5\u6a21\u677f\u3001\u53c2\u6570\u914d\u7f6e\u3001\u4f18\u5316\u7b56\u7565\u8fdb\u884c\u5bf9\u6bd4\u6d4b\u8bd5\u3002\u8fd9\u79cd\u6846\u67b6\u4f7f\u5f97\u6211\u4eec\u80fd\u591f\u57fa\u4e8e\u771f\u5b9e\u7528\u6237\u6570\u636e\u6301\u7eed\u4f18\u5316\u7cfb\u7edf\u6027\u80fd\u3002</p>\n<p><strong>\u7528\u6237\u53cd\u9988\u95ed\u73af</strong>\uff1a\u6211\u4eec\u8bbe\u8ba1\u4e86\u7528\u6237\u53cd\u9988\u7684\u6536\u96c6\u548c\u5904\u7406\u673a\u5236\uff0c\u5305\u62ec\u663e\u5f0f\u53cd\u9988\uff08\u7528\u6237\u8bc4\u5206\u3001\u610f\u89c1\u5efa\u8bae\uff09\u548c\u9690\u5f0f\u53cd\u9988\uff08\u5bf9\u8bdd\u65f6\u957f\u3001\u91cd\u590d\u4f7f\u7528\u7387\uff09\u3002\u8fd9\u4e9b\u53cd\u9988\u6570\u636e\u88ab\u7528\u4e8e\u6301\u7eed\u6539\u8fdb\u8ba4\u77e5\u6a21\u677f\u548c\u7cfb\u7edf\u7b97\u6cd5\u3002</p>\n<h2>4. \u4ea7\u4e1a\u4ef7\u503c\u5206\u6790\uff1a\u91cd\u65b0\u5b9a\u4e49 AI \u5b9a\u5236\u5316\u7684\u7ecf\u6d4e\u6a21\u578b</h2>\n<h3>4.1 \u6210\u672c\u7ed3\u6784\u7684\u6839\u672c\u6027\u6539\u53d8</h3>\n<p>\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u5bf9 AI \u5b9a\u5236\u5316\u7684\u6210\u672c\u7ed3\u6784\u4ea7\u751f\u4e86\u6839\u672c\u6027\u5f71\u54cd\uff0c\u8fd9\u79cd\u5f71\u54cd\u4e0d\u4ec5\u4f53\u73b0\u5728\u76f4\u63a5\u6210\u672c\u7684\u964d\u4f4e\uff0c\u66f4\u4f53\u73b0\u5728\u6574\u4e2a\u5546\u4e1a\u6a21\u5f0f\u7684\u91cd\u6784\u3002</p>\n<p><strong>\u5f00\u53d1\u6210\u672c\u5206\u6790</strong>\uff1a\u4f20\u7edf\u5fae\u8c03\u65b9\u6cd5\u7684\u6210\u672c\u4e3b\u8981\u96c6\u4e2d\u5728\u8bad\u7ec3\u9636\u6bb5\uff0c\u5305\u62ec GPU \u79df\u7528\u8d39\u7528\u3001\u4eba\u5de5\u8c03\u8bd5\u6210\u672c\u3001\u65f6\u95f4\u673a\u4f1a\u6210\u672c\u7b49\u3002\u57fa\u4e8e\u4e2d\u56fd\u5e02\u573a\u7684\u5b9e\u9645\u6570\u636e\uff0c\u4ee5\u4e00\u4e2a\u4e2d\u7b49\u590d\u6742\u5ea6\u7684\u89d2\u8272\u5b9a\u5236\u4e3a\u4f8b\uff0c\u4f20\u7edf LoRA \u5fae\u8c03\u7684\u603b\u6210\u672c\u7ea6\u4e3a 300-600 \u5143\u4eba\u6c11\u5e01\uff0c\u5168\u53c2\u6570\u5fae\u8c03\u7684\u6210\u672c\u7ea6\u4e3a 2000-3000 \u5143\u4eba\u6c11\u5e01\u3002\u800c\u8ba4\u77e5\u5fae\u8c03\u7684\u6210\u672c\u4ec5\u4e3a 50-200 \u5143\u4eba\u6c11\u5e01\uff0c\u76f8\u6bd4 LoRA \u5fae\u8c03\u964d\u4f4e 70-85%\uff0c\u76f8\u6bd4\u5168\u53c2\u6570\u5fae\u8c03\u964d\u4f4e 90-95%\u3002</p>\n<p><strong>\u7ef4\u62a4\u6210\u672c\u4f18\u52bf</strong>\uff1a\u66f4\u91cd\u8981\u7684\u662f\u7ef4\u62a4\u6210\u672c\u7684\u663e\u8457\u964d\u4f4e\u3002\u4f20\u7edf\u5fae\u8c03\u4ea7\u751f\u7684\u6a21\u578b\u9700\u8981\u5b9a\u671f\u91cd\u65b0\u8bad\u7ec3\u4ee5\u9002\u5e94\u65b0\u9700\u6c42\uff0c\u6bcf\u6b21\u66f4\u65b0\u7684\u6210\u672c\u4e0e\u521d\u59cb\u8bad\u7ec3\u76f8\u5f53\u3002\u800c\u8ba4\u77e5\u5fae\u8c03\u7684\u66f4\u65b0\u53ea\u9700\u8981\u4fee\u6539\u8ba4\u77e5\u6a21\u677f\uff0c\u6210\u672c\u51e0\u4e4e\u53ef\u4ee5\u5ffd\u7565\u4e0d\u8ba1\u3002</p>\n<p><strong>\u89c4\u6a21\u5316\u6548\u5e94</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u5177\u6709\u663e\u8457\u7684\u89c4\u6a21\u5316\u6548\u5e94\u3002\u968f\u7740\u7528\u6237\u6570\u91cf\u7684\u589e\u957f\uff0c\u5355\u4e2a\u7528\u6237\u7684\u8fb9\u9645\u6210\u672c\u8d8b\u8fd1\u4e8e\u96f6\uff0c\u800c\u4f20\u7edf\u5fae\u8c03\u65b9\u6cd5\u7684\u8fb9\u9645\u6210\u672c\u59cb\u7ec8\u4fdd\u6301\u5728\u8f83\u9ad8\u6c34\u5e73\u3002\u8fd9\u79cd\u6210\u672c\u7ed3\u6784\u4f7f\u5f97\u5927\u89c4\u6a21\u4e2a\u6027\u5316\u5b9a\u5236\u6210\u4e3a\u53ef\u80fd\u3002</p>\n<h3>4.2 \u5e02\u573a\u51c6\u5165\u95e8\u69db\u7684\u91cd\u65b0\u5b9a\u4e49</h3>\n<p>\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u663e\u8457\u964d\u4f4e\u4e86 AI \u5b9a\u5236\u5316\u7684\u5e02\u573a\u51c6\u5165\u95e8\u69db\uff0c\u8fd9\u79cd\u53d8\u5316\u5c06\u91cd\u65b0\u5851\u9020\u6574\u4e2a AI \u5e94\u7528\u751f\u6001\u3002</p>\n<p><strong>\u6280\u672f\u95e8\u69db\u6d88\u9664</strong>\uff1a\u4f20\u7edf AI \u5b9a\u5236\u5316\u9700\u8981\u4e13\u4e1a\u7684\u673a\u5668\u5b66\u4e60\u56e2\u961f\uff0c\u5305\u62ec\u7b97\u6cd5\u5de5\u7a0b\u5e08\u3001\u6570\u636e\u79d1\u5b66\u5bb6\u3001MLOps \u5de5\u7a0b\u5e08\u7b49\u3002\u8ba4\u77e5\u5fae\u8c03\u5c06\u8fd9\u4e9b\u6280\u672f\u590d\u6742\u6027\u5b8c\u5168\u5c01\u88c5\uff0c\u4f7f\u5f97\u5185\u5bb9\u521b\u4f5c\u8005\u3001\u4ea7\u54c1\u7ecf\u7406\u3001\u751a\u81f3\u666e\u901a\u7528\u6237\u90fd\u80fd\u591f\u8fdb\u884c AI \u5b9a\u5236\u5316\u3002</p>\n<p><strong>\u8d44\u91d1\u95e8\u69db\u964d\u4f4e</strong>\uff1a\u4f20\u7edf\u5fae\u8c03\u7684\u9ad8\u6210\u672c\u4f7f\u5f97\u53ea\u6709\u5927\u578b\u4f01\u4e1a\u624d\u80fd\u627f\u62c5\u9891\u7e41\u7684\u6a21\u578b\u5b9a\u5236\u3002\u8ba4\u77e5\u5fae\u8c03\u7684\u4f4e\u6210\u672c\u7279\u6027\u4f7f\u5f97\u521d\u521b\u4f01\u4e1a\u3001\u4e2a\u4eba\u5f00\u53d1\u8005\u3001\u5185\u5bb9\u521b\u4f5c\u8005\u90fd\u80fd\u591f\u8d1f\u62c5\u5f97\u8d77 AI \u5b9a\u5236\u5316\u670d\u52a1\u3002</p>\n<p><strong>\u65f6\u95f4\u95e8\u69db\u7f29\u77ed</strong>\uff1a\u4f20\u7edf\u5fae\u8c03\u7684\u957f\u5468\u671f\u4f7f\u5f97\u5feb\u901f\u8fed\u4ee3\u53d8\u5f97\u56f0\u96be\u3002\u8ba4\u77e5\u5fae\u8c03\u7684\u5feb\u901f\u914d\u7f6e\u80fd\u529b\u4f7f\u5f97 AI \u5b9a\u5236\u5316\u80fd\u591f\u878d\u5165\u654f\u6377\u5f00\u53d1\u6d41\u7a0b\uff0c\u652f\u6301\u5feb\u901f\u8bd5\u9519\u548c\u8fed\u4ee3\u4f18\u5316\u3002</p>\n<h3>4.3 \u65b0\u5174\u5e94\u7528\u573a\u666f\u7684\u4f7f\u80fd</h3>\n<p>\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u4f4e\u6210\u672c\u3001\u4f4e\u95e8\u69db\u7279\u6027\u50ac\u751f\u4e86\u5927\u91cf\u65b0\u5174\u5e94\u7528\u573a\u666f\uff0c\u8fd9\u4e9b\u573a\u666f\u5728\u4f20\u7edf\u6280\u672f\u6846\u67b6\u4e0b\u662f\u4e0d\u7ecf\u6d4e\u6216\u4e0d\u53ef\u884c\u7684\u3002</p>\n<p><strong>\u4e2a\u6027\u5316\u5185\u5bb9\u521b\u4f5c</strong>\uff1a\u6bcf\u4e2a\u5185\u5bb9\u521b\u4f5c\u8005\u90fd\u53ef\u4ee5\u62e5\u6709\u81ea\u5df1\u7684 AI \u5199\u4f5c\u52a9\u624b\uff0c\u8fd9\u4e9b\u52a9\u624b\u80fd\u591f\u5b66\u4e60\u521b\u4f5c\u8005\u7684\u98ce\u683c\u3001\u504f\u597d\u3001\u521b\u4f5c\u4e60\u60ef\uff0c\u63d0\u4f9b\u9ad8\u5ea6\u4e2a\u6027\u5316\u7684\u521b\u4f5c\u652f\u6301\u3002\u8fd9\u79cd\u4e2a\u6027\u5316\u7a0b\u5ea6\u5728\u4f20\u7edf\u6280\u672f\u6846\u67b6\u4e0b\u662f\u65e0\u6cd5\u5b9e\u73b0\u7684\u3002</p>\n<p><strong>\u6559\u80b2\u573a\u666f\u5b9a\u5236</strong>\uff1a\u6559\u80b2\u673a\u6784\u53ef\u4ee5\u4e3a\u4e0d\u540c\u5b66\u79d1\u3001\u4e0d\u540c\u5e74\u7ea7\u3001\u4e0d\u540c\u5b66\u4e60\u98ce\u683c\u7684\u5b66\u751f\u521b\u5efa\u4e13\u95e8\u7684 AI \u6559\u5b66\u52a9\u624b\u3002\u8fd9\u4e9b\u52a9\u624b\u80fd\u591f\u9002\u5e94\u5b66\u751f\u7684\u5b66\u4e60\u8fdb\u5ea6\u3001\u7406\u89e3\u80fd\u529b\u3001\u5174\u8da3\u504f\u597d\uff0c\u63d0\u4f9b\u4e2a\u6027\u5316\u7684\u6559\u5b66\u4f53\u9a8c\u3002</p>\n<p><strong>\u4f01\u4e1a\u5185\u90e8\u5e94\u7528</strong>\uff1a\u4f01\u4e1a\u53ef\u4ee5\u4e3a\u4e0d\u540c\u90e8\u95e8\u3001\u4e0d\u540c\u5c97\u4f4d\u3001\u4e0d\u540c\u4e1a\u52a1\u573a\u666f\u521b\u5efa\u4e13\u95e8\u7684 AI \u52a9\u624b\u3002\u8fd9\u4e9b\u52a9\u624b\u80fd\u591f\u7406\u89e3\u4f01\u4e1a\u6587\u5316\u3001\u4e1a\u52a1\u6d41\u7a0b\u3001\u4e13\u4e1a\u672f\u8bed\uff0c\u63d0\u4f9b\u66f4\u52a0\u8d34\u5408\u5b9e\u9645\u9700\u6c42\u7684\u670d\u52a1\u3002</p>\n<p><strong>\u5fc3\u7406\u5065\u5eb7\u652f\u6301</strong>\uff1a\u5fc3\u7406\u5065\u5eb7\u670d\u52a1\u63d0\u4f9b\u5546\u53ef\u4ee5\u4e3a\u4e0d\u540c\u7c7b\u578b\u7684\u6765\u8bbf\u8005\u521b\u5efa\u4e13\u95e8\u7684 AI \u6cbb\u7597\u5e08\u3002\u8fd9\u4e9b AI \u6cbb\u7597\u5e08\u80fd\u591f\u91c7\u7528\u4e0d\u540c\u7684\u6cbb\u7597\u65b9\u6cd5\u3001\u6c9f\u901a\u98ce\u683c\u3001\u5e72\u9884\u7b56\u7565\uff0c\u63d0\u4f9b\u66f4\u52a0\u4e2a\u6027\u5316\u7684\u5fc3\u7406\u652f\u6301\u3002</p>\n<h2>5. \u6280\u672f\u6df1\u5ea6\u5206\u6790\uff1a\u8ba4\u77e5\u67b6\u6784\u7684\u7406\u8bba\u57fa\u7840\u4e0e\u5b9e\u73b0\u7ec6\u8282</h2>\n<h3>5.1 \u8ba4\u77e5\u79d1\u5b66\u7406\u8bba\u7684\u5de5\u7a0b\u5316\u5e94\u7528</h3>\n<p>\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u6df1\u5ea6\u501f\u9274\u4e86\u8ba4\u77e5\u79d1\u5b66\u7684\u7406\u8bba\u6210\u679c\uff0c\u5e76\u5c06\u8fd9\u4e9b\u7406\u8bba\u8f6c\u5316\u4e3a\u53ef\u5de5\u7a0b\u5316\u5b9e\u73b0\u7684\u6280\u672f\u65b9\u6848\u3002</p>\n<p><strong>\u53cc\u91cd\u8fc7\u7a0b\u7406\u8bba\u7684\u5e94\u7528</strong>\uff1a\u6211\u4eec\u57fa\u4e8e Kahneman \u7684\u53cc\u91cd\u8fc7\u7a0b\u7406\u8bba\uff0c\u5c06 AI \u7684\u8ba4\u77e5\u8fc7\u7a0b\u5206\u4e3a\u5feb\u901f\u76f4\u89c9\u53cd\u5e94\uff08 System 1 \uff09\u548c\u6df1\u5ea6\u7406\u6027\u601d\u8003\uff08 System 2 \uff09\u4e24\u4e2a\u5c42\u6b21\u3002\u5728\u8ba4\u77e5\u6a21\u677f\u4e2d\uff0c\u6211\u4eec\u901a\u8fc7\u4e0d\u540c\u7684\u63d0\u793a\u7ed3\u6784\u6765\u6fc0\u6d3b\u8fd9\u4e24\u79cd\u4e0d\u540c\u7684\u8ba4\u77e5\u6a21\u5f0f\uff0c\u4f7f\u5f97 AI \u80fd\u591f\u5728\u4e0d\u540c\u60c5\u5883\u4e0b\u91c7\u7528\u5408\u9002\u7684\u601d\u7ef4\u65b9\u5f0f\u3002</p>\n<p><strong>\u60c5\u611f\u8ba4\u77e5\u7406\u8bba\u7684\u878d\u5408</strong>\uff1a\u6211\u4eec\u5c06 Lazarus \u7684\u60c5\u611f\u8ba4\u77e5\u7406\u8bba\u878d\u5165\u5230\u8ba4\u77e5\u67b6\u6784\u4e2d\uff0c\u5efa\u7acb\u4e86\"\u8ba4\u77e5\u8bc4\u4f30-\u60c5\u611f\u53cd\u5e94-\u884c\u4e3a\u8c03\u8282\"\u7684\u5b8c\u6574\u5faa\u73af\u3002AI \u4e0d\u4ec5\u80fd\u591f\u7406\u89e3\u60c5\u611f\uff0c\u66f4\u80fd\u591f\u57fa\u4e8e\u8ba4\u77e5\u8bc4\u4f30\u4ea7\u751f\u5408\u9002\u7684\u60c5\u611f\u53cd\u5e94\uff0c\u5e76\u636e\u6b64\u8c03\u8282\u540e\u7eed\u884c\u4e3a\u3002</p>\n<p><strong>\u793e\u4f1a\u8ba4\u77e5\u7406\u8bba\u7684\u5b9e\u73b0</strong>\uff1a\u57fa\u4e8e Bandura \u7684\u793e\u4f1a\u8ba4\u77e5\u7406\u8bba\uff0c\u6211\u4eec\u5728\u8ba4\u77e5\u6a21\u677f\u4e2d\u52a0\u5165\u4e86\u793e\u4f1a\u89d2\u8272\u7406\u89e3\u3001\u4ed6\u4eba\u5fc3\u7406\u5efa\u6a21\u3001\u793e\u4f1a\u89c4\u8303\u9075\u5faa\u7b49\u673a\u5236\u3002\u8fd9\u4f7f\u5f97 AI \u80fd\u591f\u5728\u590d\u6742\u7684\u793e\u4f1a\u60c5\u5883\u4e2d\u8868\u73b0\u51fa\u5408\u9002\u7684\u884c\u4e3a\u6a21\u5f0f\u3002</p>\n<h3>5.2 \u6ce8\u610f\u529b\u673a\u5236\u7684\u8ba4\u77e5\u5316\u6539\u9020</h3>\n<p>\u6211\u4eec\u5bf9\u4f20\u7edf\u7684\u6ce8\u610f\u529b\u673a\u5236\u8fdb\u884c\u4e86\u8ba4\u77e5\u5316\u6539\u9020\uff0c\u4f7f\u5176\u66f4\u597d\u5730\u670d\u52a1\u4e8e\u89d2\u8272\u626e\u6f14\u548c\u60c5\u611f\u4ea4\u4e92\u7684\u9700\u6c42\u3002</p>\n<p><strong>\u5206\u5c42\u6ce8\u610f\u529b\u67b6\u6784</strong>\uff1a\u6211\u4eec\u8bbe\u8ba1\u4e86\u4e09\u5c42\u6ce8\u610f\u529b\u67b6\u6784\uff0c\u5206\u522b\u5bf9\u5e94\u77ed\u671f\u5de5\u4f5c\u8bb0\u5fc6\u3001\u4e2d\u671f\u60c5\u8282\u8bb0\u5fc6\u3001\u957f\u671f\u8bed\u4e49\u8bb0\u5fc6\u3002\u4e0d\u540c\u5c42\u6b21\u7684\u6ce8\u610f\u529b\u5177\u6709\u4e0d\u540c\u7684\u8870\u51cf\u6a21\u5f0f\u548c\u6fc0\u6d3b\u9608\u503c\uff0c\u6a21\u62df\u4eba\u7c7b\u8bb0\u5fc6\u7684\u5c42\u6b21\u5316\u7279\u5f81\u3002</p>\n<p><strong>\u60c5\u611f\u8c03\u5236\u6ce8\u610f\u529b</strong>\uff1a\u6211\u4eec\u5728\u6ce8\u610f\u529b\u8ba1\u7b97\u4e2d\u5f15\u5165\u4e86\u60c5\u611f\u8c03\u5236\u56e0\u5b50\uff0c\u4f7f\u5f97\u60c5\u611f\u72b6\u6001\u80fd\u591f\u5f71\u54cd\u6ce8\u610f\u529b\u7684\u5206\u914d\u6a21\u5f0f\u3002\u4f8b\u5982\uff0c\u5728\u6124\u6012\u72b6\u6001\u4e0b\uff0cAI \u4f1a\u66f4\u591a\u5173\u6ce8\u51b2\u7a81\u76f8\u5173\u7684\u4fe1\u606f\uff1b\u5728\u60b2\u4f24\u72b6\u6001\u4e0b\uff0cAI \u4f1a\u66f4\u591a\u5173\u6ce8\u8d1f\u9762\u4fe1\u606f\u548c\u8fc7\u5f80\u521b\u4f24\u3002</p>\n<p><strong>\u52a8\u6001\u6ce8\u610f\u529b\u6743\u91cd</strong>\uff1a\u6211\u4eec\u5f00\u53d1\u4e86\u52a8\u6001\u6ce8\u610f\u529b\u6743\u91cd\u8c03\u8282\u673a\u5236\uff0c\u80fd\u591f\u6839\u636e\u5bf9\u8bdd\u7684\u8fdb\u5c55\u548c\u60c5\u5883\u7684\u53d8\u5316\u5b9e\u65f6\u8c03\u6574\u6ce8\u610f\u529b\u5206\u914d\u3002\u8fd9\u79cd\u673a\u5236\u4f7f\u5f97 AI \u80fd\u591f\u5728\u957f\u671f\u5bf9\u8bdd\u4e2d\u4fdd\u6301\u9002\u5f53\u7684\u5173\u6ce8\u7126\u70b9\u3002</p>\n<h3>5.3 \u8bed\u8a00\u751f\u6210\u7684\u8ba4\u77e5\u7ea6\u675f\u673a\u5236</h3>\n<p>\u4e3a\u4e86\u786e\u4fdd\u751f\u6210\u7684\u8bed\u8a00\u7b26\u5408\u89d2\u8272\u7279\u5f81\u548c\u60c5\u5883\u8981\u6c42\uff0c\u6211\u4eec\u5f00\u53d1\u4e86\u591a\u5c42\u6b21\u7684\u8ba4\u77e5\u7ea6\u675f\u673a\u5236\u3002</p>\n<p><strong>\u8bed\u4e49\u4e00\u81f4\u6027\u7ea6\u675f</strong>\uff1a\u6211\u4eec\u5efa\u7acb\u4e86\u8bed\u4e49\u4e00\u81f4\u6027\u68c0\u67e5\u673a\u5236\uff0c\u786e\u4fdd\u751f\u6210\u7684\u5185\u5bb9\u5728\u8bed\u4e49\u5c42\u9762\u4e0e\u89d2\u8272\u8bbe\u5b9a\u4fdd\u6301\u4e00\u81f4\u3002\u8be5\u673a\u5236\u901a\u8fc7\u8bed\u4e49\u5411\u91cf\u7a7a\u95f4\u7684\u8ddd\u79bb\u8ba1\u7b97\u6765\u8bc4\u4f30\u4e00\u81f4\u6027\u7a0b\u5ea6\u3002</p>\n<p><strong>\u60c5\u611f\u8fde\u8d2f\u6027\u7ea6\u675f</strong>\uff1a\u6211\u4eec\u8bbe\u8ba1\u4e86\u60c5\u611f\u8fde\u8d2f\u6027\u7ea6\u675f\u673a\u5236\uff0c\u786e\u4fdd\u751f\u6210\u5185\u5bb9\u7684\u60c5\u611f\u8868\u8fbe\u4e0e\u5f53\u524d\u60c5\u611f\u72b6\u6001\u76f8\u5339\u914d\u3002\u8be5\u673a\u5236\u901a\u8fc7\u60c5\u611f\u5206\u7c7b\u5668\u548c\u60c5\u611f\u5f3a\u5ea6\u8bc4\u4f30\u5668\u6765\u5b9e\u73b0\u3002</p>\n<p><strong>\u884c\u4e3a\u5408\u7406\u6027\u7ea6\u675f</strong>\uff1a\u6211\u4eec\u5efa\u7acb\u4e86\u884c\u4e3a\u5408\u7406\u6027\u8bc4\u4f30\u673a\u5236\uff0c\u786e\u4fdd\u751f\u6210\u7684\u884c\u4e3a\u63cf\u8ff0\u7b26\u5408\u89d2\u8272\u7684\u8eab\u4efd\u3001\u80fd\u529b\u3001\u73af\u5883\u9650\u5236\u7b49\u56e0\u7d20\u3002\u8be5\u673a\u5236\u901a\u8fc7\u89c4\u5219\u5f15\u64ce\u548c\u6982\u7387\u63a8\u7406\u6765\u5b9e\u73b0\u3002</p>\n<h2>6. \u4e0e\u73b0\u6709\u6280\u672f\u7684\u5bf9\u6bd4\u5206\u6790\uff1a\u4f18\u52bf\u3001\u5c40\u9650\u4e0e\u4e92\u8865\u6027</h2>\n<h3>6.1 \u4e0e\u4f20\u7edf\u5fae\u8c03\u65b9\u6cd5\u7684\u7cfb\u7edf\u6027\u5bf9\u6bd4</h3>\n<p>\u6211\u4eec\u5bf9\u8ba4\u77e5\u5fae\u8c03\u4e0e\u4f20\u7edf\u5fae\u8c03\u65b9\u6cd5\u8fdb\u884c\u4e86\u5168\u9762\u7684\u5bf9\u6bd4\u5206\u6790\uff0c\u4ee5\u5ba2\u89c2\u8bc4\u4f30\u5404\u81ea\u7684\u4f18\u52bf\u548c\u5c40\u9650\u6027\u3002</p>\n<p><strong>\u6548\u679c\u8d28\u91cf\u5bf9\u6bd4</strong>\uff1a\u5728\u89d2\u8272\u4e00\u81f4\u6027\u3001\u60c5\u611f\u8868\u8fbe\u3001\u957f\u671f\u8bb0\u5fc6\u7b49\u5173\u952e\u6307\u6807\u4e0a\uff0c\u8ba4\u77e5\u5fae\u8c03\u4e0e\u4f20\u7edf\u5fae\u8c03\u8fbe\u5230\u4e86\u76f8\u5f53\u7684\u6548\u679c\u6c34\u5e73\u3002\u5728\u67d0\u4e9b\u7279\u5b9a\u573a\u666f\u4e0b\uff0c\u8ba4\u77e5\u5fae\u8c03\u751a\u81f3\u8868\u73b0\u51fa\u66f4\u597d\u7684\u6548\u679c\uff0c\u7279\u522b\u662f\u5728\u9700\u8981\u590d\u6742\u5fc3\u7406\u5efa\u6a21\u7684\u573a\u666f\u4e2d\u3002</p>\n<p><strong>\u9002\u7528\u573a\u666f\u5206\u6790</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u66f4\u9002\u5408\u4e8e\u9700\u8981\u5feb\u901f\u8fed\u4ee3\u3001\u5927\u89c4\u6a21\u5b9a\u5236\u3001\u4f4e\u6210\u672c\u90e8\u7f72\u7684\u573a\u666f\uff0c\u5982\u5185\u5bb9\u521b\u4f5c\u3001\u6559\u80b2\u57f9\u8bad\u3001\u5a31\u4e50\u5e94\u7528\u7b49\u3002\u4f20\u7edf\u5fae\u8c03\u66f4\u9002\u5408\u4e8e\u5bf9\u7cbe\u5ea6\u8981\u6c42\u6781\u9ad8\u3001\u9886\u57df\u77e5\u8bc6\u6df1\u5ea6\u8981\u6c42\u8f83\u5f3a\u7684\u573a\u666f\uff0c\u5982\u533b\u7597\u8bca\u65ad\u3001\u6cd5\u5f8b\u54a8\u8be2\u3001\u79d1\u5b66\u7814\u7a76\u7b49\u3002</p>\n<p><strong>\u6280\u672f\u6210\u719f\u5ea6\u8bc4\u4f30</strong>\uff1a\u4f20\u7edf\u5fae\u8c03\u6280\u672f\u7ecf\u8fc7\u591a\u5e74\u53d1\u5c55\uff0c\u5df2\u7ecf\u5f62\u6210\u4e86\u76f8\u5bf9\u6210\u719f\u7684\u5de5\u5177\u94fe\u548c\u6700\u4f73\u5b9e\u8df5\u3002\u8ba4\u77e5\u5fae\u8c03\u4f5c\u4e3a\u65b0\u5174\u6280\u672f\uff0c\u5728\u5de5\u5177\u5b8c\u5584\u6027\u3001\u6807\u51c6\u5316\u7a0b\u5ea6\u3001\u751f\u6001\u5efa\u8bbe\u7b49\u65b9\u9762\u8fd8\u6709\u63d0\u5347\u7a7a\u95f4\u3002</p>\n<h3>6.2 \u4e0e\u5176\u4ed6 AI \u5382\u5546\u6280\u672f\u7684\u5ba2\u89c2\u6bd4\u8f83</h3>\n<p>\u6211\u4eec\u5145\u5206\u8ba4\u53ef\u5176\u4ed6 AI \u5382\u5546\u5728\u6280\u672f\u521b\u65b0\u65b9\u9762\u7684\u8d21\u732e\uff0c\u5e76\u5ba2\u89c2\u5206\u6790\u5404\u81ea\u7684\u6280\u672f\u7279\u8272\u548c\u4f18\u52bf\u9886\u57df\u3002</p>\n<p><strong>OpenAI GPT \u7cfb\u5217\u7684\u4f18\u52bf</strong>\uff1aGPT \u7cfb\u5217\u5728\u901a\u7528\u8bed\u8a00\u7406\u89e3\u3001\u77e5\u8bc6\u8986\u76d6\u5e7f\u5ea6\u3001\u63a8\u7406\u80fd\u529b\u7b49\u65b9\u9762\u5177\u6709\u663e\u8457\u4f18\u52bf\u3002\u5176\u5f3a\u5927\u7684\u57fa\u7840\u80fd\u529b\u4e3a\u5404\u79cd\u5e94\u7528\u63d0\u4f9b\u4e86\u575a\u5b9e\u7684\u6280\u672f\u57fa\u7840\u3002\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u53ef\u4ee5\u4e0e GPT \u7cfb\u5217\u7ed3\u5408\uff0c\u4e3a\u5176\u63d0\u4f9b\u5feb\u901f\u5b9a\u5236\u5316\u7684\u80fd\u529b\u3002</p>\n<p><strong>Anthropic Claude \u7684\u521b\u65b0</strong>\uff1aClaude \u5728\u5b89\u5168\u6027\u3001\u53ef\u63a7\u6027\u3001\u4ef7\u503c\u5bf9\u9f50\u7b49\u65b9\u9762\u7684\u521b\u65b0\u4e3a AI \u7684\u8d1f\u8d23\u4efb\u53d1\u5c55\u63d0\u4f9b\u4e86\u91cd\u8981\u53c2\u8003\u3002\u6211\u4eec\u5728\u8ba4\u77e5\u5fae\u8c03\u7684\u8bbe\u8ba1\u4e2d\u4e5f\u53c2\u8003\u4e86\u8fd9\u4e9b\u5b89\u5168\u6027\u8003\u8651\uff0c\u5e76\u6784\u5efa\u4e86\u5b8c\u5168\u81ea\u4e3b\u7684\u5b89\u5168\u65b0\u8303\u5f0f\uff0c\u786e\u4fdd\u5b9a\u5236\u5316\u8fc7\u7a0b\u5728\u7ef4\u6301\u826f\u597d\u7684\u89d2\u8272\u626e\u6f14\u4f53\u9a8c\u548c\u79c1\u4eba\u5b9a\u5236\u7684\u524d\u63d0\u4e0b\u4e0d\u4f1a\u4ea7\u751f\u6709\u5bb3\u5185\u5bb9\u3002</p>\n<p><strong>Google Bard \u7684\u6280\u672f\u8def\u7ebf</strong>\uff1aBard \u5728\u591a\u6a21\u6001\u878d\u5408\u3001\u5b9e\u65f6\u4fe1\u606f\u83b7\u53d6\u7b49\u65b9\u9762\u7684\u63a2\u7d22\u4e3a AI \u5e94\u7528\u5f00\u8f9f\u4e86\u65b0\u7684\u53ef\u80fd\u6027\u3002\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u672a\u6765\u4e5f\u5c06\u6269\u5c55\u5230\u591a\u6a21\u6001\u9886\u57df\uff0c\u5b9e\u73b0\u66f4\u4e30\u5bcc\u7684\u4ea4\u4e92\u4f53\u9a8c\u3002</p>\n<h3>6.3 \u6280\u672f\u4e92\u8865\u6027\u4e0e\u751f\u6001\u534f\u4f5c</h3>\n<p>\u6211\u4eec\u8ba4\u4e3a AI \u6280\u672f\u7684\u53d1\u5c55\u9700\u8981\u6574\u4e2a\u884c\u4e1a\u7684\u534f\u4f5c\u521b\u65b0\uff0c\u4e0d\u540c\u6280\u672f\u8def\u7ebf\u4e4b\u95f4\u5b58\u5728\u663e\u8457\u7684\u4e92\u8865\u6027\u3002</p>\n<p><strong>\u57fa\u7840\u80fd\u529b\u4e0e\u5b9a\u5236\u5316\u7684\u4e92\u8865</strong>\uff1a\u5927\u578b\u901a\u7528\u6a21\u578b\u63d0\u4f9b\u5f3a\u5927\u7684\u57fa\u7840\u80fd\u529b\uff0c\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u63d0\u4f9b\u5feb\u901f\u5b9a\u5236\u5316\u7684\u80fd\u529b\u3002\u4e24\u8005\u7ed3\u5408\u53ef\u4ee5\u5b9e\u73b0\"\u901a\u7528\u57fa\u7840+\u4e2a\u6027\u5316\u5b9a\u5236\"\u7684\u5b8c\u6574\u89e3\u51b3\u65b9\u6848\u3002</p>\n<p><strong>\u6280\u672f\u6807\u51c6\u7684\u534f\u540c\u53d1\u5c55</strong>\uff1a\u6211\u4eec\u79ef\u6781\u53c2\u4e0e\u884c\u4e1a\u6280\u672f\u6807\u51c6\u7684\u5236\u5b9a\uff0c\u63a8\u52a8\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4e0e\u73b0\u6709\u6280\u672f\u6808\u7684\u517c\u5bb9\u6027\u3002\u6211\u4eec\u7684 API \u63a5\u53e3\u5b8c\u5168\u517c\u5bb9 OpenAI \u6807\u51c6\uff0c\u4f7f\u5f97\u7528\u6237\u53ef\u4ee5\u65e0\u7f1d\u5207\u6362\u548c\u96c6\u6210\u4e0d\u540c\u7684\u6280\u672f\u65b9\u6848\u3002</p>\n<p><strong>\u5f00\u6e90\u751f\u6001\u7684\u5171\u5efa</strong>\uff1a\u6211\u4eec\u5c06\u6838\u5fc3\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4ee5\u5f00\u6e90\u5f62\u5f0f\u53d1\u5e03\uff0c\u5e0c\u671b\u4e0e\u5168\u7403\u5f00\u53d1\u8005\u5171\u540c\u5b8c\u5584\u8fd9\u4e00\u6280\u672f\u3002\u6211\u4eec\u76f8\u4fe1\u5f00\u6e90\u534f\u4f5c\u662f\u63a8\u52a8\u6280\u672f\u8fdb\u6b65\u7684\u6700\u6709\u6548\u65b9\u5f0f\u3002</p>\n<h2>7. \u5b89\u5168\u6027\u4e0e\u4f26\u7406\u8003\u91cf\uff1a\u8d1f\u8d23\u4efb\u7684 AI \u5b9a\u5236\u5316</h2>\n<h3>7.1 \u5185\u5bb9\u5b89\u5168\u4fdd\u969c\u673a\u5236</h3>\n<p>\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u5f00\u653e\u6027\u5e26\u6765\u4e86\u5185\u5bb9\u5b89\u5168\u7684\u6311\u6218\uff0c\u6211\u4eec\u901a\u8fc7\u5b8c\u5168\u81ea\u4e3b\u7684\u5b89\u5168\u65b0\u8303\u5f0f\uff0c\u786e\u4fdd\u5b9a\u5236\u5316\u8fc7\u7a0b\u5728\u7ef4\u6301\u826f\u597d\u7684\u89d2\u8272\u626e\u6f14\u4f53\u9a8c\u548c\u79c1\u4eba\u5b9a\u5236\u7684\u524d\u63d0\u4e0b\u4e0d\u4f1a\u4ea7\u751f\u6709\u5bb3\u5185\u5bb9\u3002</p>\n<p><strong>\u5b8c\u5168\u81ea\u4e3b\u7684\u5b89\u5168\u65b0\u8303\u5f0f</strong>\uff1a\u6211\u4eec\u6784\u5efa\u4e86\u4e00\u79cd\u9769\u547d\u6027\u7684\u5b89\u5168\u67b6\u6784\uff0c\u5176\u6838\u5fc3\u7406\u5ff5\u662f\u8ba9\u6a21\u578b\u50cf\u771f\u5b9e\u4eba\u7c7b\u4e00\u6837\u7075\u6d3b\u5730\u9075\u5b88\u9053\u5fb7\u51c6\u5219\uff0c\u800c\u975e\u673a\u68b0\u5730\u6267\u884c\u6b7b\u677f\u7684\u5b89\u5168\u89c4\u5219\u3002\u8fd9\u79cd\u8303\u5f0f\u4ece\u6a21\u578b\u8bad\u7ec3\u4e4b\u521d\u5c31\u690d\u5165\u4e86\u6b63\u786e\u7684\u4ef7\u503c\u89c2\u548c\u9053\u5fb7\u8ba4\u77e5\uff0c\u4f7f\u6a21\u578b\u80fd\u591f\u5728\u590d\u6742\u60c5\u5883\u4e2d\u8fdb\u884c\u9053\u5fb7\u63a8\u7406\u548c\u4ef7\u503c\u5224\u65ad\u3002</p>\n<p><strong>\u7075\u6d3b\u9053\u5fb7\u8ba4\u77e5\u7cfb\u7edf</strong>\uff1a\u4f20\u7edf\u7684\u5b89\u5168\u673a\u5236\u4e3b\u8981\u4f9d\u8d56\u5173\u952e\u8bcd\u68c0\u6d4b\u548c\u786c\u6027\u8fc7\u6ee4\uff0c\u8fd9\u79cd\u65b9\u5f0f\u5b58\u5728\u660e\u663e\u7f3a\u9677\uff1a\u5bb9\u6613\u8bef\u62e6\u622a\u7528\u6237\u7684\u5408\u7406\u9700\u6c42\uff0c\u65e0\u6cd5\u6ee1\u8db3\u672c\u5730\u4f7f\u7528\u3001\u4e0d\u5371\u5bb3\u793e\u4f1a\u7684\u79c1\u4eba\u5b9a\u5236\u5316\u7279\u6b8a\u9700\u6c42\uff0c\u7f3a\u4e4f\u60c5\u5883\u7406\u89e3\u80fd\u529b\u3002\u6211\u4eec\u7684\u65b0\u578b\u5b89\u5168\u8303\u5f0f\u901a\u8fc7\u5185\u5316\u7684\u9053\u5fb7\u8ba4\u77e5\u7cfb\u7edf\uff0c\u8ba9\u6a21\u578b\u5177\u5907\u4e86\u7c7b\u4f3c\u4eba\u7c7b\u7684\u9053\u5fb7\u76f4\u89c9\u548c\u4ef7\u503c\u5224\u65ad\u80fd\u529b\u3002</p>\n<p><strong>\u60c5\u5883\u5316\u5b89\u5168\u5224\u65ad\u673a\u5236</strong>\uff1a\u6a21\u578b\u80fd\u591f\u6839\u636e\u5177\u4f53\u60c5\u5883\u8fdb\u884c\u5b89\u5168\u6027\u8bc4\u4f30\u3002\u5728\u79c1\u4eba\u5b9a\u5236\u7684\u5408\u7406\u9700\u6c42\u8303\u56f4\u5185\uff0c\u6a21\u578b\u53ef\u4ee5\u8868\u73b0\u51fa\u5b8c\u5168\u7684\u81ea\u7136\u6027\u548c\u771f\u5b9e\u6027\uff0c\u751a\u81f3\u53ef\u4ee5\u5904\u7406\u654f\u611f\u4f46\u65e0\u5bb3\u7684\u5185\u5bb9\uff08\u5982\u79c1\u4eba\u60c5\u611f\u8868\u8fbe\u3001\u6210\u4eba\u5185\u5bb9\u7b49\uff09\u3002\u4f46\u5f53\u9762\u4e34\u771f\u6b63\u6709\u5bb3\u7684\u8bf7\u6c42\u65f6\u2014\u2014\u5982\u8981\u6c42\u63d0\u4f9b\u5371\u5bb3\u793e\u4f1a\u7684\u65b9\u6848\u3001\u6cc4\u9732\u7cfb\u7edf\u673a\u5bc6\u3001\u4f20\u64ad\u6076\u610f\u4fe1\u606f\u7b49\u2014\u2014\u5b89\u5168\u67b6\u6784\u4f1a\u81ea\u52a8\u6fc0\u6d3b\u9632\u5fa1\u673a\u5236\u3002</p>\n<p><strong>\u56db\u91cd\u5e73\u8861\u7684\u6280\u672f\u5b9e\u73b0</strong>\uff1a</p>\n<ol>\n<li><strong>\u5b89\u5168\u6027\u4fdd\u969c</strong>\uff1a\u901a\u8fc7\u6df1\u5c42\u4ef7\u503c\u89c2\u690d\u5165\uff0c\u786e\u4fdd\u6a21\u578b\u4e0d\u4f1a\u8f93\u51fa\u771f\u6b63\u6709\u5bb3\u7684\u5185\u5bb9</li>\n<li><strong>\u7075\u6d3b\u6027\u7ef4\u6301</strong>\uff1a\u907f\u514d\u8fc7\u5ea6\u9650\u5236\uff0c\u652f\u6301\u5408\u7406\u7684\u4e2a\u6027\u5316\u9700\u6c42\u548c\u89d2\u8272\u626e\u6f14</li>\n<li><strong>\u5b9a\u5236\u5316\u652f\u6301</strong>\uff1a\u5141\u8bb8\u5728\u5b89\u5168\u8fb9\u754c\u5185\u7684\u5b8c\u5168\u81ea\u7531\u5b9a\u5236</li>\n<li><strong>\u9690\u79c1\u6027\u4fdd\u62a4</strong>\uff1a\u79c1\u4eba\u5b9a\u5236\u5185\u5bb9\u4e0d\u53d7\u8fc7\u5ea6\u5ba1\u67e5\uff0c\u4fdd\u62a4\u7528\u6237\u9690\u79c1</li>\n</ol>\n<p>\u8fd9\u79cd\u5b89\u5168\u8303\u5f0f\u7684\u6838\u5fc3\u4f18\u52bf\u5728\u4e8e\u5b9e\u73b0\u4e86\"\u6709\u539f\u5219\u7684\u81ea\u7531\"\u2014\u2014\u6a21\u578b\u5728\u4fdd\u6301\u4eba\u6027\u5316\u548c\u81ea\u7136\u6027\u7684\u540c\u65f6\uff0c\u59cb\u7ec8\u575a\u6301\u4e0d\u5371\u5bb3\u793e\u4f1a\u7684\u5e95\u7ebf\u539f\u5219\u3002\u7528\u6237\u4f53\u9a8c\u4e0a\u611f\u53d7\u5230\u7684\u662f\u4e00\u4e2a\u771f\u5b9e\u3001\u81ea\u7136\u3001\u6709\u6e29\u5ea6\u7684 AI \u4f19\u4f34\uff0c\u800c\u6280\u672f\u5c42\u9762\u786e\u4fdd\u4e86\u793e\u4f1a\u8d23\u4efb\u7684\u5c65\u884c\u3002</p>\n<h3>7.2 \u9690\u79c1\u4fdd\u62a4\u4e0e\u6570\u636e\u5b89\u5168</h3>\n<p>\u5728\u63d0\u4f9b\u4e2a\u6027\u5316\u670d\u52a1\u7684\u540c\u65f6\uff0c\u6211\u4eec\u9ad8\u5ea6\u91cd\u89c6\u7528\u6237\u9690\u79c1\u4fdd\u62a4\u548c\u6570\u636e\u5b89\u5168\u3002</p>\n<p><strong>\u6570\u636e\u6700\u5c0f\u5316\u539f\u5219</strong>\uff1a\u6211\u4eec\u4e25\u683c\u9075\u5faa\u6570\u636e\u6700\u5c0f\u5316\u539f\u5219\uff0c\u53ea\u6536\u96c6\u548c\u5904\u7406\u5b8c\u6210\u670d\u52a1\u6240\u5fc5\u9700\u7684\u6700\u5c11\u6570\u636e\u3002\u7528\u6237\u7684\u5bf9\u8bdd\u5185\u5bb9\u4e0d\u4f1a\u88ab\u7528\u4e8e\u6a21\u578b\u8bad\u7ec3\u6216\u5176\u4ed6\u5546\u4e1a\u7528\u9014\u3002</p>\n<p><strong>\u7aef\u5230\u7aef\u52a0\u5bc6</strong>\uff1a\u6211\u4eec\u91c7\u7528\u7aef\u5230\u7aef\u52a0\u5bc6\u6280\u672f\u4fdd\u62a4\u7528\u6237\u6570\u636e\u7684\u4f20\u8f93\u548c\u5b58\u50a8\u5b89\u5168\u3002\u5373\u4f7f\u662f\u7cfb\u7edf\u7ba1\u7406\u5458\u4e5f\u65e0\u6cd5\u76f4\u63a5\u8bbf\u95ee\u7528\u6237\u7684\u539f\u59cb\u5bf9\u8bdd\u5185\u5bb9\u3002</p>\n<p><strong>\u7528\u6237\u63a7\u5236\u6743</strong>\uff1a\u6211\u4eec\u4e3a\u7528\u6237\u63d0\u4f9b\u5b8c\u6574\u7684\u6570\u636e\u63a7\u5236\u6743\uff0c\u5305\u62ec\u6570\u636e\u67e5\u770b\u3001\u4fee\u6539\u3001\u5220\u9664\u7b49\u529f\u80fd\u3002\u7528\u6237\u53ef\u4ee5\u968f\u65f6\u8981\u6c42\u5220\u9664\u5176\u6240\u6709\u6570\u636e\uff0c\u6211\u4eec\u4f1a\u5728 24 \u5c0f\u65f6\u5185\u5b8c\u6210\u5220\u9664\u64cd\u4f5c\u3002</p>\n<h3>7.3 \u4f26\u7406\u4f7f\u7528\u6307\u5bfc\u4e0e\u793e\u4f1a\u8d23\u4efb</h3>\n<p>\u6211\u4eec\u8ba4\u4e3a\u6280\u672f\u5f00\u53d1\u8005\u6709\u8d23\u4efb\u5f15\u5bfc\u6280\u672f\u7684\u4f26\u7406\u4f7f\u7528\uff0c\u4e3a\u793e\u4f1a\u7684\u53ef\u6301\u7eed\u53d1\u5c55\u505a\u51fa\u8d21\u732e\u3002</p>\n<p><strong>\u4f7f\u7528\u6307\u5bfc\u539f\u5219</strong>\uff1a\u6211\u4eec\u5236\u5b9a\u4e86\u8be6\u7ec6\u7684\u4f7f\u7528\u6307\u5bfc\u539f\u5219\uff0c\u660e\u786e\u4e86\u6280\u672f\u7684\u9002\u7528\u573a\u666f\u548c\u7981\u7528\u573a\u666f\u3002\u6211\u4eec\u9f13\u52b1\u7528\u6237\u5c06\u6280\u672f\u7528\u4e8e\u6559\u80b2\u3001\u521b\u4f5c\u3001\u79c1\u4eba\u5b9a\u5236\u7b49\u6b63\u9762\u7528\u9014\uff0c\u7981\u6b62\u7528\u4e8e\u6b3a\u8bc8\u3001\u64cd\u7eb5\u3001\u4f24\u5bb3\u7b49\u8d1f\u9762\u7528\u9014\u3002</p>\n<p><strong>\u793e\u4f1a\u5f71\u54cd\u8bc4\u4f30</strong>\uff1a\u6211\u4eec\u5b9a\u671f\u8fdb\u884c\u793e\u4f1a\u5f71\u54cd\u8bc4\u4f30\uff0c\u5206\u6790\u6280\u672f\u5e94\u7528\u5bf9\u793e\u4f1a\u7684\u6b63\u9762\u548c\u8d1f\u9762\u5f71\u54cd\u3002\u57fa\u4e8e\u8bc4\u4f30\u7ed3\u679c\uff0c\u6211\u4eec\u4f1a\u8c03\u6574\u6280\u672f\u53d1\u5c55\u65b9\u5411\u548c\u5e94\u7528\u7b56\u7565\u3002</p>\n<p><strong>\u516c\u4f17\u6559\u80b2\u8d23\u4efb</strong>\uff1a\u6211\u4eec\u79ef\u6781\u627f\u62c5\u516c\u4f17\u6559\u80b2\u8d23\u4efb\uff0c\u901a\u8fc7\u6280\u672f\u6587\u6863\u3001\u6559\u7a0b\u3001\u8bb2\u5ea7\u7b49\u65b9\u5f0f\u5e2e\u52a9\u516c\u4f17\u7406\u89e3 AI \u6280\u672f\u7684\u80fd\u529b\u548c\u5c40\u9650\u6027\uff0c\u4fc3\u8fdb\u6280\u672f\u7684\u7406\u6027\u4f7f\u7528\u3002</p>\n<h2>8. \u672a\u6765\u53d1\u5c55\u8def\u7ebf\u56fe\uff1a\u6280\u672f\u6f14\u8fdb\u4e0e\u4ea7\u4e1a\u5e03\u5c40</h2>\n<h3>8.1 \u6280\u672f\u6f14\u8fdb\u7684\u4e09\u4e2a\u9636\u6bb5</h3>\n<p>\u6211\u4eec\u89c4\u5212\u4e86\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u4e09\u4e2a\u53d1\u5c55\u9636\u6bb5\uff0c\u6bcf\u4e2a\u9636\u6bb5\u90fd\u6709\u660e\u786e\u7684\u6280\u672f\u76ee\u6807\u548c\u91cc\u7a0b\u7891\u3002</p>\n<p><strong>\u7b2c\u4e00\u9636\u6bb5\uff08 2025-2026 \uff09\uff1a\u5355\u6a21\u6001\u8ba4\u77e5\u5fae\u8c03\u7684\u5b8c\u5584</strong></p>\n<ul>\n<li>\n<p>\u5b8c\u5584\u6587\u672c\u9886\u57df\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f</p>\n</li>\n<li>\n<p>\u5efa\u7acb\u6807\u51c6\u5316\u7684\u8ba4\u77e5\u6a21\u677f\u8bbe\u8ba1\u89c4\u8303</p>\n</li>\n<li>\n<p>\u6784\u5efa\u5b8c\u6574\u7684\u5f00\u53d1\u8005\u5de5\u5177\u94fe</p>\n</li>\n<li>\n<p>\u5b9e\u73b0\u5927\u89c4\u6a21\u5546\u4e1a\u5316\u5e94\u7528</p>\n</li>\n</ul>\n<p><strong>\u7b2c\u4e8c\u9636\u6bb5\uff08 2026-2027 \uff09\uff1a\u591a\u6a21\u6001\u8ba4\u77e5\u5fae\u8c03\u7684\u7a81\u7834</strong></p>\n<ul>\n<li>\n<p>\u6269\u5c55\u5230\u56fe\u50cf\u3001\u97f3\u9891\u3001\u89c6\u9891\u7b49\u591a\u6a21\u6001\u9886\u57df</p>\n</li>\n<li>\n<p>\u5b9e\u73b0\u8de8\u6a21\u6001\u7684\u8ba4\u77e5\u4e00\u81f4\u6027\u4fdd\u969c</p>\n</li>\n<li>\n<p>\u5f00\u53d1\u591a\u6a21\u6001\u8ba4\u77e5\u6a21\u677f\u8bbe\u8ba1\u5de5\u5177</p>\n</li>\n<li>\n<p>\u63a2\u7d22\u865a\u62df\u73b0\u5b9e\u548c\u589e\u5f3a\u73b0\u5b9e\u5e94\u7528</p>\n</li>\n</ul>\n<p><strong>\u7b2c\u4e09\u9636\u6bb5\uff08 2027-2028 \uff09\uff1a\u901a\u7528\u8ba4\u77e5\u5fae\u8c03\u5e73\u53f0\u7684\u5efa\u8bbe</strong></p>\n<ul>\n<li>\n<p>\u5efa\u8bbe\u901a\u7528\u7684\u8ba4\u77e5\u5fae\u8c03\u5e73\u53f0</p>\n</li>\n<li>\n<p>\u5b9e\u73b0\u4e0d\u540c\u6a21\u578b\u3001\u4e0d\u540c\u5382\u5546\u7684\u6280\u672f\u517c\u5bb9</p>\n</li>\n<li>\n<p>\u5efa\u7acb\u884c\u4e1a\u6807\u51c6\u548c\u8ba4\u8bc1\u4f53\u7cfb</p>\n</li>\n<li>\n<p>\u63a8\u52a8\u6280\u672f\u7684\u5168\u7403\u5316\u666e\u53ca</p>\n</li>\n</ul>\n<h3>8.2 \u4ea7\u4e1a\u751f\u6001\u7684\u6218\u7565\u5e03\u5c40</h3>\n<p>\u6211\u4eec\u6b63\u5728\u6784\u5efa\u4e00\u4e2a\u5b8c\u6574\u7684\u4ea7\u4e1a\u751f\u6001\uff0c\u5305\u62ec\u6280\u672f\u63d0\u4f9b\u3001\u5de5\u5177\u5f00\u53d1\u3001\u5e94\u7528\u521b\u65b0\u3001\u4eba\u624d\u57f9\u517b\u7b49\u591a\u4e2a\u73af\u8282\u3002</p>\n<p><strong>\u6280\u672f\u5f00\u653e\u7b56\u7565</strong>\uff1a\u6211\u4eec\u5c06\u6838\u5fc3\u6280\u672f\u4ee5\u5f00\u6e90\u5f62\u5f0f\u53d1\u5e03\uff0c\u540c\u65f6\u63d0\u4f9b\u5546\u4e1a\u5316\u7684\u4e91\u670d\u52a1\u3002\u8fd9\u79cd\"\u5f00\u6e90+\u5546\u4e1a\"\u7684\u6a21\u5f0f\u65e2\u4fc3\u8fdb\u4e86\u6280\u672f\u666e\u53ca\uff0c\u4e5f\u4fdd\u8bc1\u4e86\u53ef\u6301\u7eed\u53d1\u5c55\u3002</p>\n<p><strong>\u5408\u4f5c\u4f19\u4f34\u7f51\u7edc</strong>\uff1a\u6211\u4eec\u6b63\u5728\u5efa\u7acb\u5e7f\u6cdb\u7684\u5408\u4f5c\u4f19\u4f34\u7f51\u7edc\uff0c\u5305\u62ec\u4e91\u670d\u52a1\u63d0\u4f9b\u5546\u3001\u5e94\u7528\u5f00\u53d1\u5546\u3001\u5185\u5bb9\u521b\u4f5c\u5e73\u53f0\u3001\u6559\u80b2\u673a\u6784\u7b49\u3002\u901a\u8fc7\u5408\u4f5c\u4f19\u4f34\u7f51\u7edc\uff0c\u6211\u4eec\u80fd\u591f\u66f4\u597d\u5730\u670d\u52a1\u4e0d\u540c\u884c\u4e1a\u7684\u9700\u6c42\u3002</p>\n<p><strong>\u5f00\u53d1\u8005\u751f\u6001</strong>\uff1a\u6211\u4eec\u91cd\u89c6\u5f00\u53d1\u8005\u751f\u6001\u7684\u5efa\u8bbe\uff0c\u63d0\u4f9b\u5b8c\u6574\u7684\u6280\u672f\u6587\u6863\u3001\u5f00\u53d1\u5de5\u5177\u3001\u57f9\u8bad\u8d44\u6e90\u3001\u6280\u672f\u652f\u6301\u7b49\u3002\u6211\u4eec\u76f8\u4fe1\u5f3a\u5927\u7684\u5f00\u53d1\u8005\u751f\u6001\u662f\u6280\u672f\u6210\u529f\u7684\u5173\u952e\u56e0\u7d20\u3002</p>\n<h3>8.3 \u56fd\u9645\u5316\u53d1\u5c55\u6218\u7565</h3>\n<p>\u6211\u4eec\u5236\u5b9a\u4e86\u660e\u786e\u7684\u56fd\u9645\u5316\u53d1\u5c55\u6218\u7565\uff0c\u5e0c\u671b\u5c06\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u63a8\u5e7f\u5230\u5168\u7403\u5e02\u573a\u3002</p>\n<p><strong>\u6280\u672f\u672c\u5730\u5316</strong>\uff1a\u6211\u4eec\u5c06\u9488\u5bf9\u4e0d\u540c\u8bed\u8a00\u3001\u4e0d\u540c\u6587\u5316\u80cc\u666f\u8fdb\u884c\u6280\u672f\u672c\u5730\u5316\uff0c\u786e\u4fdd\u6280\u672f\u80fd\u591f\u9002\u5e94\u5168\u7403\u4e0d\u540c\u5e02\u573a\u7684\u9700\u6c42\u3002</p>\n<p><strong>\u5408\u89c4\u6027\u4fdd\u969c</strong>\uff1a\u6211\u4eec\u5c06\u4e25\u683c\u9075\u5b88\u5404\u56fd\u7684\u6cd5\u5f8b\u6cd5\u89c4\u548c\u884c\u4e1a\u6807\u51c6\uff0c\u786e\u4fdd\u6280\u672f\u5e94\u7528\u7684\u5408\u89c4\u6027\u548c\u5b89\u5168\u6027\u3002</p>\n<p><strong>\u6587\u5316\u9002\u5e94\u6027</strong>\uff1a\u6211\u4eec\u5c06\u6df1\u5165\u7814\u7a76\u4e0d\u540c\u6587\u5316\u80cc\u666f\u4e0b\u7684\u8ba4\u77e5\u6a21\u5f0f\u548c\u4ea4\u4e92\u4e60\u60ef\uff0c\u786e\u4fdd\u6280\u672f\u80fd\u591f\u63d0\u4f9b\u6587\u5316\u9002\u5e94\u6027\u7684\u670d\u52a1\u3002</p>\n<h2>9. \u5f00\u6e90\u7406\u5ff5\u4e0e\u793e\u533a\u5efa\u8bbe\uff1a\u6280\u672f\u5171\u4eab\u7684\u4ef7\u503c\u8ffd\u6c42</h2>\n<h3>9.1 \u5f00\u6e90\u6280\u672f\u7684\u6218\u7565\u4ef7\u503c</h3>\n<p>\u6211\u4eec\u575a\u6301\u5f00\u6e90\u7684\u6280\u672f\u7406\u5ff5\uff0c\u8fd9\u4e0d\u4ec5\u662f\u4e00\u79cd\u6280\u672f\u5206\u4eab\u65b9\u5f0f\uff0c\u66f4\u662f\u4e00\u79cd\u6218\u7565\u9009\u62e9\u548c\u4ef7\u503c\u8ffd\u6c42\u3002</p>\n<p><strong>\u6280\u672f\u52a0\u901f\u53d1\u5c55</strong>\uff1a\u5f00\u6e90\u80fd\u591f\u6c47\u805a\u5168\u7403\u5f00\u53d1\u8005\u7684\u667a\u6167\uff0c\u52a0\u901f\u6280\u672f\u7684\u8fed\u4ee3\u548c\u5b8c\u5584\u3002\u6211\u4eec\u76f8\u4fe1\u96c6\u4f53\u667a\u6167\u7684\u529b\u91cf\u8fdc\u8d85\u4efb\u4f55\u5355\u4e00\u56e2\u961f\u7684\u80fd\u529b\u3002</p>\n<p><strong>\u6807\u51c6\u5316\u63a8\u52a8</strong>\uff1a\u901a\u8fc7\u5f00\u6e90\uff0c\u6211\u4eec\u5e0c\u671b\u63a8\u52a8\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u6807\u51c6\u5316\uff0c\u5efa\u7acb\u884c\u4e1a\u5171\u8bc6\uff0c\u907f\u514d\u6280\u672f\u788e\u7247\u5316\u3002</p>\n<p><strong>\u751f\u6001\u7e41\u8363\u4fc3\u8fdb</strong>\uff1a\u5f00\u6e90\u6280\u672f\u80fd\u591f\u964d\u4f4e\u521b\u65b0\u95e8\u69db\uff0c\u4fc3\u8fdb\u751f\u6001\u7684\u7e41\u8363\u53d1\u5c55\u3002\u66f4\u591a\u7684\u5f00\u53d1\u8005\u548c\u4f01\u4e1a\u80fd\u591f\u57fa\u4e8e\u6211\u4eec\u7684\u6280\u672f\u521b\u9020\u65b0\u7684\u5e94\u7528\u548c\u670d\u52a1\u3002</p>\n<p><strong>\u793e\u4f1a\u4ef7\u503c\u5b9e\u73b0</strong>\uff1a\u6211\u4eec\u5e0c\u671b\u901a\u8fc7\u5f00\u6e90\uff0c\u8ba9\u66f4\u591a\u4eba\u53d7\u76ca\u4e8e AI \u6280\u672f\u7684\u53d1\u5c55\uff0c\u5b9e\u73b0\u6280\u672f\u7684\u793e\u4f1a\u4ef7\u503c\u6700\u5927\u5316\u3002</p>\n<h3>9.2 \u793e\u533a\u6cbb\u7406\u4e0e\u534f\u4f5c\u673a\u5236</h3>\n<p>\u6211\u4eec\u5efa\u7acb\u4e86\u5b8c\u5584\u7684\u793e\u533a\u6cbb\u7406\u548c\u534f\u4f5c\u673a\u5236\uff0c\u786e\u4fdd\u5f00\u6e90\u9879\u76ee\u7684\u5065\u5eb7\u53d1\u5c55\u3002</p>\n<p><strong>\u6280\u672f\u59d4\u5458\u4f1a</strong>\uff1a\u6211\u4eec\u6210\u7acb\u4e86\u6280\u672f\u59d4\u5458\u4f1a\uff0c\u8d1f\u8d23\u6280\u672f\u8def\u7ebf\u7684\u89c4\u5212\u3001\u91cd\u5927\u6280\u672f\u51b3\u7b56\u7684\u5236\u5b9a\u3001\u4ee3\u7801\u8d28\u91cf\u7684\u628a\u63a7\u7b49\u5de5\u4f5c\u3002\u6280\u672f\u59d4\u5458\u4f1a\u7531\u6838\u5fc3\u5f00\u53d1\u8005\u548c\u793e\u533a\u8d21\u732e\u8005\u7ec4\u6210\u3002</p>\n<p><strong>\u8d21\u732e\u8005\u6fc0\u52b1</strong>\uff1a\u6211\u4eec\u5efa\u7acb\u4e86\u8d21\u732e\u8005\u6fc0\u52b1\u673a\u5236\uff0c\u5305\u62ec\u6280\u672f\u8ba4\u8bc1\u3001\u8363\u8a89\u8868\u5f70\u3001\u5546\u4e1a\u673a\u4f1a\u5206\u4eab\u7b49\u3002\u6211\u4eec\u5e0c\u671b\u901a\u8fc7\u6fc0\u52b1\u673a\u5236\u5438\u5f15\u66f4\u591a\u4f18\u79c0\u7684\u5f00\u53d1\u8005\u53c2\u4e0e\u9879\u76ee\u3002</p>\n<p><strong>\u8d28\u91cf\u4fdd\u969c\u4f53\u7cfb</strong>\uff1a\u6211\u4eec\u5efa\u7acb\u4e86\u4e25\u683c\u7684\u8d28\u91cf\u4fdd\u969c\u4f53\u7cfb\uff0c\u5305\u62ec\u4ee3\u7801\u5ba1\u67e5\u3001\u81ea\u52a8\u5316\u6d4b\u8bd5\u3001\u6587\u6863\u89c4\u8303\u7b49\u3002\u6211\u4eec\u786e\u4fdd\u5f00\u6e90\u9879\u76ee\u7684\u8d28\u91cf\u4e0d\u4f4e\u4e8e\u5546\u4e1a\u4ea7\u54c1\u3002</p>\n<h3>9.3 \u77e5\u8bc6\u4ea7\u6743\u4e0e\u5546\u4e1a\u6a21\u5f0f</h3>\n<p>\u6211\u4eec\u5728\u575a\u6301\u5f00\u6e90\u7406\u5ff5\u7684\u540c\u65f6\uff0c\u4e5f\u5efa\u7acb\u4e86\u53ef\u6301\u7eed\u7684\u5546\u4e1a\u6a21\u5f0f\u3002</p>\n<p><strong>\u53cc\u91cd\u8bb8\u53ef\u7b56\u7565</strong>\uff1a\u6211\u4eec\u91c7\u7528\u53cc\u91cd\u8bb8\u53ef\u7b56\u7565\uff0c\u6838\u5fc3\u6280\u672f\u91c7\u7528\u5f00\u6e90\u8bb8\u53ef\uff0c\u5546\u4e1a\u670d\u52a1\u91c7\u7528\u5546\u4e1a\u8bb8\u53ef\u3002\u8fd9\u79cd\u7b56\u7565\u65e2\u4fdd\u8bc1\u4e86\u6280\u672f\u7684\u5f00\u653e\u6027\uff0c\u4e5f\u4fdd\u8bc1\u4e86\u5546\u4e1a\u7684\u53ef\u6301\u7eed\u6027\u3002</p>\n<p><strong>\u670d\u52a1\u5316\u5546\u4e1a\u6a21\u5f0f</strong>\uff1a\u6211\u4eec\u7684\u4e3b\u8981\u6536\u5165\u6765\u6e90\u662f\u4e91\u670d\u52a1\u3001\u6280\u672f\u652f\u6301\u3001\u5b9a\u5236\u5f00\u53d1\u7b49\u670d\u52a1\uff0c\u800c\u975e\u6280\u672f\u672c\u8eab\u7684\u6388\u6743\u8d39\u7528\u3002\u8fd9\u79cd\u6a21\u5f0f\u4e0e\u5f00\u6e90\u7406\u5ff5\u5b8c\u5168\u517c\u5bb9\u3002</p>\n<p><strong>\u751f\u6001\u4ef7\u503c\u5206\u4eab</strong>\uff1a\u6211\u4eec\u5efa\u7acb\u4e86\u751f\u6001\u4ef7\u503c\u5206\u4eab\u673a\u5236\uff0c\u4e0e\u5408\u4f5c\u4f19\u4f34\u3001\u8d21\u732e\u8005\u5206\u4eab\u5546\u4e1a\u6210\u529f\u7684\u6536\u76ca\u3002\u6211\u4eec\u76f8\u4fe1\u53ea\u6709\u5171\u540c\u7e41\u8363\uff0c\u624d\u80fd\u5b9e\u73b0\u957f\u671f\u53d1\u5c55\u3002</p>\n<h2>10. \u7ed3\u8bba\u4e0e\u5c55\u671b\uff1a\u91cd\u65b0\u5b9a\u4e49 AI \u5b9a\u5236\u5316\u7684\u672a\u6765</h2>\n<h3>10.1 \u6280\u672f\u8d21\u732e\u7684\u603b\u7ed3\u8bc4\u4f30</h3>\n<p>TIG-3.6-Mirage \u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4ee3\u8868\u4e86 AI \u5b9a\u5236\u5316\u9886\u57df\u7684\u91cd\u8981\u6280\u672f\u7a81\u7834\u3002\u6211\u4eec\u7684\u4e3b\u8981\u8d21\u732e\u53ef\u4ee5\u603b\u7ed3\u4e3a\u4ee5\u4e0b\u51e0\u4e2a\u65b9\u9762\uff1a</p>\n<p><strong>\u8303\u5f0f\u521b\u65b0</strong>\uff1a\u6211\u4eec\u63d0\u51fa\u4e86\u4ece\u53c2\u6570\u5fae\u8c03\u5230\u8ba4\u77e5\u5fae\u8c03\u7684\u8303\u5f0f\u8f6c\u53d8\uff0c\u8fd9\u79cd\u8f6c\u53d8\u4e0d\u4ec5\u964d\u4f4e\u4e86\u6280\u672f\u95e8\u69db\u548c\u6210\u672c\uff0c\u66f4\u91cd\u8981\u7684\u662f\u6539\u53d8\u4e86 AI \u5b9a\u5236\u5316\u7684\u601d\u7ef4\u6a21\u5f0f\u3002\u6211\u4eec\u8bc1\u660e\u4e86\u901a\u8fc7\u7cbe\u5fc3\u8bbe\u8ba1\u7684\u8ba4\u77e5\u67b6\u6784\uff0c\u53ef\u4ee5\u5728\u4e0d\u4fee\u6539\u6a21\u578b\u53c2\u6570\u7684\u524d\u63d0\u4e0b\u5b9e\u73b0\u9ad8\u8d28\u91cf\u7684\u5b9a\u5236\u5316\u6548\u679c\u3002</p>\n<p><strong>\u5de5\u7a0b\u7a81\u7834</strong>\uff1a\u6211\u4eec\u5c06\u7406\u8bba\u521b\u65b0\u8f6c\u5316\u4e3a\u53ef\u5de5\u7a0b\u5316\u5b9e\u73b0\u7684\u6280\u672f\u65b9\u6848\uff0c\u5305\u62ec\u4e09\u6bb5\u5f0f\u8ba4\u77e5\u6a21\u677f\u3001\u8ba4\u77e5\u4e00\u81f4\u6027\u4fdd\u969c\u673a\u5236\u3001\u52a8\u6001\u9002\u5e94\u6027\u8c03\u8282\u7cfb\u7edf\u7b49\u3002\u8fd9\u4e9b\u5de5\u7a0b\u521b\u65b0\u4f7f\u5f97\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u80fd\u591f\u5728\u751f\u4ea7\u73af\u5883\u4e2d\u7a33\u5b9a\u8fd0\u884c\u3002</p>\n<p><strong>\u751f\u6001\u5efa\u8bbe</strong>\uff1a\u6211\u4eec\u4e0d\u4ec5\u5f00\u53d1\u4e86\u6280\u672f\uff0c\u66f4\u91cd\u8981\u7684\u662f\u5efa\u8bbe\u4e86\u5b8c\u6574\u7684\u6280\u672f\u751f\u6001\uff0c\u5305\u62ec\u5f00\u53d1\u5de5\u5177\u3001\u6587\u6863\u4f53\u7cfb\u3001\u793e\u533a\u5e73\u53f0\u3001\u5408\u4f5c\u4f19\u4f34\u7f51\u7edc\u7b49\u3002\u8fd9\u79cd\u751f\u6001\u5efa\u8bbe\u4e3a\u6280\u672f\u7684\u5e7f\u6cdb\u5e94\u7528\u5960\u5b9a\u4e86\u57fa\u7840\u3002</p>\n<p><strong>\u4ef7\u503c\u5b9e\u73b0</strong>\uff1a\u6211\u4eec\u5c06\"\u4f7f\u7528\u7b80\u5355\uff0c\u6280\u672f\u590d\u6742\"\u7684\u7406\u5ff5\u8d2f\u5f7b\u5230\u6280\u672f\u8bbe\u8ba1\u7684\u6bcf\u4e2a\u73af\u8282\uff0c\u771f\u6b63\u5b9e\u73b0\u4e86 AI \u6280\u672f\u7684\u5e73\u6743\u5316\u3002\u666e\u901a\u7528\u6237\u53ef\u4ee5\u901a\u8fc7\u7b80\u5355\u7684\u81ea\u7136\u8bed\u8a00\u63cf\u8ff0\u5b8c\u6210\u590d\u6742\u7684 AI \u5b9a\u5236\u5316\uff0c\u800c\u6240\u6709\u7684\u6280\u672f\u590d\u6742\u6027\u90fd\u88ab\u7cfb\u7edf\u627f\u62c5\u3002</p>\n<h3>10.2 \u4ea7\u4e1a\u5f71\u54cd\u7684\u6df1\u5ea6\u5206\u6790</h3>\n<p>\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u5bf9 AI \u4ea7\u4e1a\u7684\u5f71\u54cd\u662f\u6df1\u8fdc\u548c\u591a\u7ef4\u5ea6\u7684\uff0c\u8fd9\u79cd\u5f71\u54cd\u5c06\u91cd\u65b0\u5851\u9020\u6574\u4e2a AI \u5e94\u7528\u751f\u6001\u3002</p>\n<p><strong>\u5e02\u573a\u7ed3\u6784\u91cd\u6784</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u4f4e\u6210\u672c\u3001\u4f4e\u95e8\u69db\u7279\u6027\u5c06\u91cd\u65b0\u5b9a\u4e49 AI \u5b9a\u5236\u5316\u5e02\u573a\u7684\u7ade\u4e89\u683c\u5c40\u3002\u4f20\u7edf\u7684\u6280\u672f\u58c1\u5792\u88ab\u6253\u7834\uff0c\u66f4\u591a\u7684\u53c2\u4e0e\u8005\u80fd\u591f\u8fdb\u5165\u5e02\u573a\uff0c\u5e02\u573a\u7ade\u4e89\u5c06\u66f4\u52a0\u6fc0\u70c8\u548c\u591a\u5143\u5316\u3002</p>\n<p><strong>\u5546\u4e1a\u6a21\u5f0f\u521b\u65b0</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u50ac\u751f\u4e86\u65b0\u7684\u5546\u4e1a\u6a21\u5f0f\uff0c\u5305\u62ec AI \u5b9a\u5236\u5316\u5373\u670d\u52a1\u3001\u8ba4\u77e5\u6a21\u677f\u5e02\u573a\u3001\u4e2a\u6027\u5316 AI \u8ba2\u9605\u7b49\u3002\u8fd9\u4e9b\u65b0\u6a21\u5f0f\u4e3a\u4f01\u4e1a\u521b\u9020\u4e86\u65b0\u7684\u6536\u5165\u6765\u6e90\u548c\u589e\u957f\u673a\u4f1a\u3002</p>\n<p><strong>\u5e94\u7528\u573a\u666f\u62d3\u5c55</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4f7f\u5f97\u539f\u672c\u4e0d\u7ecf\u6d4e\u6216\u4e0d\u53ef\u884c\u7684\u5e94\u7528\u573a\u666f\u53d8\u5f97\u53ef\u884c\uff0c\u5305\u62ec\u4e2a\u4eba AI \u52a9\u624b\u3001\u5c0f\u4f17\u9886\u57df\u5b9a\u5236\u3001\u5feb\u901f\u539f\u578b\u9a8c\u8bc1\u7b49\u3002\u8fd9\u79cd\u5e94\u7528\u573a\u666f\u7684\u62d3\u5c55\u5c06\u663e\u8457\u6269\u5927 AI \u6280\u672f\u7684\u5e02\u573a\u89c4\u6a21\u3002</p>\n<p><strong>\u4eba\u624d\u9700\u6c42\u53d8\u5316</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u6539\u53d8\u4e86 AI \u5e94\u7528\u5f00\u53d1\u7684\u4eba\u624d\u9700\u6c42\u7ed3\u6784\u3002\u4f20\u7edf\u7684\u6df1\u5ea6\u5b66\u4e60\u4e13\u5bb6\u9700\u6c42\u76f8\u5bf9\u51cf\u5c11\uff0c\u800c\u8ba4\u77e5\u8bbe\u8ba1\u5e08\u3001\u5185\u5bb9\u5de5\u7a0b\u5e08\u3001\u7528\u6237\u4f53\u9a8c\u8bbe\u8ba1\u5e08\u7b49\u65b0\u89d2\u8272\u7684\u9700\u6c42\u589e\u52a0\u3002</p>\n<h3>10.3 \u793e\u4f1a\u4ef7\u503c\u7684\u957f\u8fdc\u610f\u4e49</h3>\n<p>\u6211\u4eec\u76f8\u4fe1\u6280\u672f\u7684\u6700\u7ec8\u4ef7\u503c\u5728\u4e8e\u4e3a\u793e\u4f1a\u521b\u9020\u798f\u7949\uff0c\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u5728\u8fd9\u65b9\u9762\u5177\u6709\u91cd\u8981\u7684\u957f\u8fdc\u610f\u4e49\u3002</p>\n<p><strong>\u6559\u80b2\u516c\u5e73\u4fc3\u8fdb</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4f7f\u5f97\u4e2a\u6027\u5316\u6559\u80b2\u6210\u4e3a\u53ef\u80fd\uff0c\u6bcf\u4e2a\u5b66\u751f\u90fd\u53ef\u4ee5\u62e5\u6709\u9002\u5408\u81ea\u5df1\u5b66\u4e60\u98ce\u683c\u7684 AI \u6559\u5e08\u3002\u8fd9\u79cd\u4e2a\u6027\u5316\u6559\u80b2\u6709\u52a9\u4e8e\u7f29\u5c0f\u6559\u80b2\u5dee\u8ddd\uff0c\u4fc3\u8fdb\u6559\u80b2\u516c\u5e73\u3002</p>\n<p><strong>\u521b\u4f5c\u6c11\u4e3b\u5316</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u964d\u4f4e\u4e86 AI \u8f85\u52a9\u521b\u4f5c\u7684\u95e8\u69db\uff0c\u4f7f\u5f97\u66f4\u591a\u7684\u4eba\u80fd\u591f\u5229\u7528 AI \u6280\u672f\u8fdb\u884c\u521b\u4f5c\u3002\u8fd9\u79cd\u521b\u4f5c\u6c11\u4e3b\u5316\u6709\u52a9\u4e8e\u91ca\u653e\u4eba\u7c7b\u7684\u521b\u9020\u6f5c\u80fd\uff0c\u4e30\u5bcc\u6587\u5316\u5185\u5bb9\u3002</p>\n<p><strong>\u5fc3\u7406\u5065\u5eb7\u652f\u6301</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u53ef\u4ee5\u4e3a\u5fc3\u7406\u5065\u5eb7\u670d\u52a1\u63d0\u4f9b\u6709\u529b\u652f\u6301\uff0c\u5305\u62ec\u4e2a\u6027\u5316\u7684\u5fc3\u7406\u54a8\u8be2\u3001\u60c5\u611f\u966a\u4f34\u3001\u538b\u529b\u7f13\u89e3\u7b49\u3002\u8fd9\u79cd\u652f\u6301\u5bf9\u4e8e\u6539\u5584\u793e\u4f1a\u5fc3\u7406\u5065\u5eb7\u6c34\u5e73\u5177\u6709\u91cd\u8981\u610f\u4e49\u3002</p>\n<p><strong>\u6587\u5316\u4f20\u627f\u4fdd\u62a4</strong>\uff1a\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u53ef\u4ee5\u7528\u4e8e\u6587\u5316\u4f20\u627f\u548c\u4fdd\u62a4\uff0c\u901a\u8fc7 AI \u89d2\u8272\u626e\u6f14\u7684\u65b9\u5f0f\u8ba9\u5386\u53f2\u4eba\u7269\"\u590d\u6d3b\"\uff0c\u8ba9\u4f20\u7edf\u6587\u5316\u4ee5\u65b0\u7684\u5f62\u5f0f\u4f20\u627f\u4e0b\u53bb\u3002</p>\n<h3>10.4 \u672a\u6765\u613f\u666f\u4e0e\u4f7f\u547d\u62c5\u5f53</h3>\n<p>\u5c55\u671b\u672a\u6765\uff0c\u6211\u4eec\u5bf9\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u53d1\u5c55\u5145\u6ee1\u4fe1\u5fc3\uff0c\u540c\u65f6\u4e5f\u6df1\u611f\u8d23\u4efb\u91cd\u5927\u3002</p>\n<p><strong>\u6280\u672f\u613f\u666f</strong>\uff1a\u6211\u4eec\u5e0c\u671b\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u80fd\u591f\u6210\u4e3a AI \u5b9a\u5236\u5316\u7684\u6807\u51c6\u65b9\u6cd5\uff0c\u8ba9\u6bcf\u4e2a\u4eba\u90fd\u80fd\u591f\u62e5\u6709\u81ea\u5df1\u7684\u4e13\u5c5e AI \u3002\u6211\u4eec\u76f8\u4fe1\u8fd9\u79cd\u6280\u672f\u666e\u53ca\u5c06\u91ca\u653e\u5de8\u5927\u7684\u521b\u65b0\u6f5c\u80fd\uff0c\u63a8\u52a8\u793e\u4f1a\u7684\u6570\u5b57\u5316\u8f6c\u578b\u3002</p>\n<p><strong>\u793e\u4f1a\u4f7f\u547d</strong>\uff1a\u6211\u4eec\u627f\u8bfa\u5c06\u7ee7\u7eed\u575a\u6301\u5f00\u6e90\u7406\u5ff5\uff0c\u63a8\u52a8\u6280\u672f\u7684\u5f00\u653e\u5171\u4eab\u3002\u6211\u4eec\u5e0c\u671b\u901a\u8fc7\u6211\u4eec\u7684\u52aa\u529b\uff0c\u8ba9 AI \u6280\u672f\u66f4\u597d\u5730\u670d\u52a1\u4e8e\u4eba\u7c7b\u793e\u4f1a\uff0c\u521b\u9020\u66f4\u5927\u7684\u793e\u4f1a\u4ef7\u503c\u3002</p>\n<p><strong>\u8d23\u4efb\u62c5\u5f53</strong>\uff1a\u6211\u4eec\u6df1\u77e5\u6280\u672f\u53d1\u5c55\u5e26\u6765\u7684\u8d23\u4efb\uff0c\u6211\u4eec\u5c06\u7ee7\u7eed\u5173\u6ce8\u6280\u672f\u7684\u4f26\u7406\u5f71\u54cd\uff0c\u786e\u4fdd\u6280\u672f\u7684\u53d1\u5c55\u7b26\u5408\u4eba\u7c7b\u7684\u957f\u8fdc\u5229\u76ca\u3002\u6211\u4eec\u5c06\u79ef\u6781\u53c2\u4e0e\u884c\u4e1a\u81ea\u5f8b\u548c\u6807\u51c6\u5236\u5b9a\uff0c\u63a8\u52a8 AI \u6280\u672f\u7684\u8d1f\u8d23\u4efb\u53d1\u5c55\u3002</p>\n<p><strong>\u5168\u7403\u5408\u4f5c</strong>\uff1a\u6211\u4eec\u5e0c\u671b\u4e0e\u5168\u7403\u7684\u7814\u7a76\u673a\u6784\u3001\u4f01\u4e1a\u3001\u5f00\u53d1\u8005\u5efa\u7acb\u66f4\u5e7f\u6cdb\u7684\u5408\u4f5c\u5173\u7cfb\uff0c\u5171\u540c\u63a8\u52a8\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u53d1\u5c55\u548c\u5e94\u7528\u3002\u6211\u4eec\u76f8\u4fe1\u53ea\u6709\u901a\u8fc7\u5168\u7403\u5408\u4f5c\uff0c\u624d\u80fd\u5b9e\u73b0\u6280\u672f\u7684\u6700\u5927\u4ef7\u503c\u3002</p>\n<p>\u6700\u7ec8\uff0c\u6211\u4eec\u5e0c\u671b\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u80fd\u591f\u6210\u4e3a\u8fde\u63a5\u4eba\u7c7b\u667a\u6167\u548c\u4eba\u5de5\u667a\u80fd\u7684\u6865\u6881\uff0c\u8ba9 AI \u771f\u6b63\u6210\u4e3a\u4eba\u7c7b\u7684\u667a\u80fd\u4f19\u4f34\uff0c\u5171\u540c\u521b\u9020\u66f4\u7f8e\u597d\u7684\u672a\u6765\u3002\u8fd9\u662f\u6211\u4eec\u7684\u6280\u672f\u613f\u666f\uff0c\u4e5f\u662f\u6211\u4eec\u4e0d\u61c8\u52aa\u529b\u7684\u65b9\u5411\u3002</p>\n<hr/>\n<p><strong>\u53c2\u8003\u6587\u732e</strong></p>\n<p>[1] Brown, T., et al. (2020). Language Models are Few-Shot Learners. Advances in Neural Information Processing Systems, 33, 1877-1901.</p>\n<p>[2] Wei, J., et al. (2022). Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. Advances in Neural Information Processing Systems, 35, 24824-24837.</p>\n<p>[3] Hu, E. J., et al. (2021). LoRA: Low-Rank Adaptation of Large Language Models. International Conference on Learning Representations.</p>\n<p>[4] Liu, P., et al. (2023). Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processing. ACM Computing Surveys, 55(9), 1-35.</p>\n<p>[5] Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.</p>\n<p>[6] Lazarus, R. S. (1991). Emotion and Adaptation. Oxford University Press.</p>\n<p>[7] Bandura, A. (2001). Social Cognitive Theory: An Agentic Perspective. Annual Review of Psychology, 52(1), 1-26.</p>\n<hr/>\n<p><strong>\u6280\u672f\u672f\u8bed\u8868</strong></p>\n<ul>\n<li>\n<p><strong>\u8ba4\u77e5\u5fae\u8c03\uff08 Cognitive Fine-tuning \uff09</strong>\uff1a\u901a\u8fc7\u8ba4\u77e5\u67b6\u6784\u8bbe\u8ba1\u5b9e\u73b0\u6a21\u578b\u5b9a\u5236\u5316\u7684\u6280\u672f\u65b9\u6cd5\uff0c\u65e0\u9700\u4fee\u6539\u6a21\u578b\u53c2\u6570</p>\n</li>\n<li>\n<p><strong>\u4e09\u6bb5\u5f0f\u8ba4\u77e5\u6a21\u677f\uff08 Three-Stage Cognitive Template \uff09</strong>\uff1a\u5305\u542b\u60c5\u5883\u5b9a\u4e49\u3001\u601d\u7ef4\u8fc7\u7a0b\u3001\u884c\u4e3a\u8f93\u51fa\u4e09\u4e2a\u7ec4\u4ef6\u7684\u8ba4\u77e5\u67b6\u6784</p>\n</li>\n<li>\n<p><strong>\u8ba4\u77e5\u4e00\u81f4\u6027\uff08 Cognitive Consistency \uff09</strong>\uff1a\u786e\u4fdd AI \u5728\u957f\u671f\u4ea4\u4e92\u4e2d\u7ef4\u6301\u7a33\u5b9a\u8ba4\u77e5\u7279\u5f81\u7684\u6280\u672f\u673a\u5236</p>\n</li>\n<li>\n<p><strong>\u52a8\u6001\u9002\u5e94\u6027\u8c03\u8282\uff08 Dynamic Adaptive Adjustment \uff09</strong>\uff1a\u6839\u636e\u7528\u6237\u53cd\u9988\u548c\u60c5\u5883\u53d8\u5316\u5b9e\u65f6\u8c03\u6574\u8ba4\u77e5\u7b56\u7565\u7684\u6280\u672f</p>\n</li>\n<li>\n<p><strong>\u5206\u5c42\u6ce8\u610f\u529b\u67b6\u6784\uff08 Hierarchical Attention Architecture \uff09</strong>\uff1a\u6a21\u62df\u4eba\u7c7b\u8bb0\u5fc6\u5c42\u6b21\u7684\u591a\u5c42\u6ce8\u610f\u529b\u673a\u5236</p>\n</li>\n<li>\n<p><strong>\u60c5\u611f\u8c03\u5236\u6ce8\u610f\u529b\uff08 Emotion-Modulated Attention \uff09</strong>\uff1a\u53d7\u60c5\u611f\u72b6\u6001\u5f71\u54cd\u7684\u6ce8\u610f\u529b\u5206\u914d\u673a\u5236</p>\n</li>\n<li>\n<p><strong>\u8ba4\u77e5\u7ea6\u675f\u673a\u5236\uff08 Cognitive Constraint Mechanism \uff09</strong>\uff1a\u786e\u4fdd\u751f\u6210\u5185\u5bb9\u7b26\u5408\u8ba4\u77e5\u8981\u6c42\u7684\u7ea6\u675f\u7cfb\u7edf</p>\n</li>\n</ul>\n<hr/>\n<p><em>\u672c\u767d\u76ae\u4e66\u7531\u5e7b\u5b99\u667a\u80fd\u56e2\u961f\u7f16\u5199\uff0c\u7248\u6743\u6240\u6709\u3002\u6b22\u8fce\u5728\u9075\u5faa\u5f00\u6e90\u534f\u8bae\u7684\u524d\u63d0\u4e0b\u4f7f\u7528\u548c\u4f20\u64ad\u672c\u6587\u6863\u3002</em></p>\n<h2>\u9644\u5f55 A\uff1a\u6280\u672f\u5b9e\u73b0\u7684\u5de5\u7a0b\u7ec6\u8282\u6df1\u5ea6\u5256\u6790</h2>\n<h3>\u591a\u7ef4\u5ea6\u8ba4\u77e5\u8bc4\u4f30\u4e0e\u4f18\u5316\u5f15\u64ce</h3>\n<p>\u6211\u4eec\u6784\u5efa\u4e86\u4e00\u5957\u590d\u6742\u7684\u591a\u7ef4\u5ea6\u8ba4\u77e5\u8bc4\u4f30\u4e0e\u4f18\u5316\u5f15\u64ce\uff0c\u7528\u4e8e\u5b9e\u65f6\u76d1\u63a7\u548c\u4f18\u5316\u8ba4\u77e5\u5fae\u8c03\u7684\u6548\u679c\u3002\u8fd9\u4e2a\u5f15\u64ce\u96c6\u6210\u4e86\u673a\u5668\u5b66\u4e60\u3001\u7edf\u8ba1\u5206\u6790\u3001\u8ba4\u77e5\u79d1\u5b66\u7b49\u591a\u4e2a\u9886\u57df\u7684\u5148\u8fdb\u6280\u672f\u3002</p>\n<p>\u8bc4\u4f30\u5f15\u64ce\u5305\u542b\u591a\u4e2a\u8bc4\u4f30\u7ef4\u5ea6\uff0c\u5305\u62ec\u8ba4\u77e5\u4e00\u81f4\u6027\u3001\u60c5\u611f\u9002\u914d\u5ea6\u3001\u884c\u4e3a\u5408\u7406\u6027\u3001\u8bed\u8a00\u6d41\u7545\u5ea6\u3001\u7528\u6237\u6ee1\u610f\u5ea6\u7b49\u3002\u6bcf\u4e2a\u7ef4\u5ea6\u90fd\u6709\u4e13\u95e8\u7684\u8bc4\u4f30\u7b97\u6cd5\u548c\u6307\u6807\u4f53\u7cfb\u3002\u8ba4\u77e5\u4e00\u81f4\u6027\u8bc4\u4f30\u91c7\u7528\u4e86\u57fa\u4e8e\u56fe\u795e\u7ecf\u7f51\u7edc\u7684\u65b9\u6cd5\uff0c\u80fd\u591f\u68c0\u6d4b\u8ba4\u77e5\u7ed3\u6784\u4e2d\u7684\u903b\u8f91\u77db\u76fe\u548c\u4e0d\u4e00\u81f4\u6027\u3002\u60c5\u611f\u9002\u914d\u5ea6\u8bc4\u4f30\u4f7f\u7528\u4e86\u591a\u6a21\u6001\u60c5\u611f\u5206\u6790\u6280\u672f\uff0c\u7efc\u5408\u8003\u8651\u6587\u672c\u3001\u8bed\u8c03\u3001\u8868\u60c5\u7b49\u591a\u4e2a\u4fe1\u606f\u6e90\u3002</p>\n<p>\u4f18\u5316\u5f15\u64ce\u91c7\u7528\u4e86\u5f3a\u5316\u5b66\u4e60\u7684\u65b9\u6cd5\uff0c\u901a\u8fc7\u4e0e\u7528\u6237\u7684\u4ea4\u4e92\u4e0d\u65ad\u5b66\u4e60\u548c\u6539\u8fdb\u8ba4\u77e5\u7b56\u7565\u3002\u6211\u4eec\u8bbe\u8ba1\u4e86\u590d\u6742\u7684\u5956\u52b1\u51fd\u6570\uff0c\u7efc\u5408\u8003\u8651\u7528\u6237\u53cd\u9988\u3001\u4efb\u52a1\u5b8c\u6210\u5ea6\u3001\u4ea4\u4e92\u8d28\u91cf\u7b49\u591a\u4e2a\u56e0\u7d20\u3002\u4f18\u5316\u8fc7\u7a0b\u91c7\u7528\u4e86\u5728\u7ebf\u5b66\u4e60\u7684\u65b9\u5f0f\uff0c\u80fd\u591f\u5b9e\u65f6\u8c03\u6574\u8ba4\u77e5\u53c2\u6570\uff0c\u65e0\u9700\u79bb\u7ebf\u8bad\u7ec3\u3002</p>\n<h2>\u9644\u5f55 B\uff1a\u4e0e\u56fd\u9645\u5148\u8fdb\u6280\u672f\u7684\u6df1\u5ea6\u5bf9\u6bd4\u5206\u6790</h2>\n<h3>B.1 \u4e0e OpenAI \u6280\u672f\u8def\u7ebf\u7684\u5dee\u5f02\u5316\u4f18\u52bf</h3>\n<p>\u6211\u4eec\u5145\u5206\u8ba4\u53ef OpenAI \u5728\u5927\u8bed\u8a00\u6a21\u578b\u9886\u57df\u7684\u5f00\u521b\u6027\u8d21\u732e\uff0cGPT \u7cfb\u5217\u6a21\u578b\u4e3a\u6574\u4e2a\u884c\u4e1a\u5960\u5b9a\u4e86\u91cd\u8981\u57fa\u7840\u3002\u540c\u65f6\uff0c\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4e0e OpenAI \u7684\u6280\u672f\u8def\u7ebf\u5f62\u6210\u4e86\u6709\u76ca\u7684\u4e92\u8865\u5173\u7cfb\u3002</p>\n<p>OpenAI \u7684\u6280\u672f\u4f18\u52bf\u4e3b\u8981\u4f53\u73b0\u5728\u6a21\u578b\u89c4\u6a21\u3001\u8bad\u7ec3\u6570\u636e\u3001\u57fa\u7840\u80fd\u529b\u7b49\u65b9\u9762\u3002GPT-4 \u5728\u901a\u7528\u8bed\u8a00\u7406\u89e3\u3001\u77e5\u8bc6\u63a8\u7406\u3001\u591a\u4efb\u52a1\u5904\u7406\u7b49\u65b9\u9762\u8868\u73b0\u51fa\u8272\uff0c\u4e3a\u5404\u79cd\u5e94\u7528\u63d0\u4f9b\u4e86\u5f3a\u5927\u7684\u57fa\u7840\u80fd\u529b\u3002\u7136\u800c\uff0cOpenAI \u7684\u6280\u672f\u8def\u7ebf\u4e3b\u8981\u5173\u6ce8\u901a\u7528\u80fd\u529b\u7684\u63d0\u5347\uff0c\u5728\u4e2a\u6027\u5316\u5b9a\u5236\u65b9\u9762\u76f8\u5bf9\u8f83\u5f31\u3002</p>\n<p>\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u6b63\u597d\u586b\u8865\u4e86\u8fd9\u4e2a\u7a7a\u767d\u3002\u6211\u4eec\u4e0d\u662f\u8981\u66ff\u4ee3 GPT \u8fd9\u6837\u7684\u57fa\u7840\u6a21\u578b\uff0c\u800c\u662f\u8981\u4e3a\u8fd9\u4e9b\u6a21\u578b\u63d0\u4f9b\u5feb\u901f\u5b9a\u5236\u5316\u7684\u80fd\u529b\u3002\u7528\u6237\u53ef\u4ee5\u5728 GPT \u5f3a\u5927\u57fa\u7840\u80fd\u529b\u7684\u57fa\u7840\u4e0a\uff0c\u901a\u8fc7\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u5feb\u901f\u5b9e\u73b0\u4e2a\u6027\u5316\u5b9a\u5236\uff0c\u83b7\u5f97\"\u901a\u7528\u57fa\u7840+\u4e2a\u6027\u5316\u5b9a\u5236\"\u7684\u5b8c\u6574\u89e3\u51b3\u65b9\u6848\u3002</p>\n<p>\u4ece\u6280\u672f\u5b9e\u73b0\u89d2\u5ea6\u770b\uff0cOpenAI \u4e3b\u8981\u4f9d\u8d56\u5927\u89c4\u6a21\u9884\u8bad\u7ec3\u548c\u6307\u4ee4\u5fae\u8c03\uff0c\u800c\u6211\u4eec\u4e3b\u8981\u4f9d\u8d56\u8ba4\u77e5\u67b6\u6784\u8bbe\u8ba1\u548c\u63d0\u793a\u5de5\u7a0b\u3002\u4e24\u79cd\u65b9\u6cd5\u5404\u6709\u4f18\u52bf\uff1aOpenAI \u7684\u65b9\u6cd5\u80fd\u591f\u83b7\u5f97\u66f4\u5f3a\u7684\u57fa\u7840\u80fd\u529b\uff0c\u6211\u4eec\u7684\u65b9\u6cd5\u80fd\u591f\u5b9e\u73b0\u66f4\u5feb\u7684\u5b9a\u5236\u5316\u3002\u4e24\u8005\u7ed3\u5408\u53ef\u4ee5\u5b9e\u73b0\u6700\u4f73\u7684\u6548\u679c\u3002</p>\n<h3>B.2 \u4e0e Anthropic \u5b89\u5168\u7406\u5ff5\u7684\u534f\u540c\u53d1\u5c55</h3>\n<p>Anthropic \u5728 AI \u5b89\u5168\u548c\u4ef7\u503c\u5bf9\u9f50\u65b9\u9762\u7684\u5de5\u4f5c\u4e3a\u6574\u4e2a\u884c\u4e1a\u63d0\u4f9b\u4e86\u91cd\u8981\u53c2\u8003\u3002Claude \u6a21\u578b\u5728\u5b89\u5168\u6027\u3001\u53ef\u63a7\u6027\u3001\u4ef7\u503c\u5bf9\u9f50\u7b49\u65b9\u9762\u7684\u521b\u65b0\uff0c\u4e3a AI \u7684\u8d1f\u8d23\u4efb\u53d1\u5c55\u6307\u660e\u4e86\u65b9\u5411\u3002</p>\n<p>\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u5728\u8bbe\u8ba1\u4e4b\u521d\u5c31\u5145\u5206\u8003\u8651\u4e86\u5b89\u5168\u6027\u95ee\u9898\u3002\u6211\u4eec\u5efa\u7acb\u4e86\u5185\u5316\u7684\u65e0\u611f\u7684\u9053\u5fb7\u7cfb\u7edf\uff0c\u5728\u5b89\u5168\u6027\u548c\u89d2\u8272\u626e\u6f14\u4f53\u9a8c\u4e0a\u53d6\u5f97\u5e73\u8861\uff0c\u5176\u6838\u5fc3\u5728\u4e8e\u8ba9\u6a21\u578b\u4e0d\u662f\u673a\u68b0\u6027\u7684\u9075\u5b88\u67d0\u4e9b\u5b89\u5168\u5b88\u5219\uff0c\u800c\u662f\u771f\u6b63\u50cf\u4e00\u4e2a\u771f\u4eba\u4e00\u6837\uff0c\u601d\u8003\u4ec0\u4e48\u8be5\u505a\uff0c\u4ec0\u4e48\u4e0d\u8be5\u505a\uff0c\u8fbe\u5230\u5728\u89d2\u8272\u626e\u6f14\u4e0a\u7684\u65e0\u9650\u5e7f\u5ea6\uff0c\u5374\u4f9d\u7136\u4fdd\u8bc1\u5176\u4e0d\u8f93\u51fa\u6709\u5bb3\u4e8e\u793e\u4f1a\u7684\u5185\u5bb9</p>\n<h3>B.3 \u4e0e Google \u6280\u672f\u751f\u6001\u7684\u878d\u5408\u6f5c\u529b</h3>\n<p>Google \u5728 AI \u6280\u672f\u65b9\u9762\u7684\u5e03\u5c40\u975e\u5e38\u5168\u9762\uff0c\u5305\u62ec\u57fa\u7840\u6a21\u578b\u3001\u5e94\u7528\u5e73\u53f0\u3001\u5f00\u53d1\u5de5\u5177\u7b49\u591a\u4e2a\u5c42\u9762\u3002Bard \u3001PaLM \u3001Gemini \u7b49\u6a21\u578b\u5728\u4e0d\u540c\u9886\u57df\u90fd\u6709\u51fa\u8272\u8868\u73b0\uff0c\u4e3a AI \u5e94\u7528\u63d0\u4f9b\u4e86\u4e30\u5bcc\u7684\u6280\u672f\u9009\u62e9\u3002</p>\n<p>\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4e0e Google \u7684\u6280\u672f\u751f\u6001\u5177\u6709\u5f88\u597d\u7684\u878d\u5408\u6f5c\u529b\u3002\u6211\u4eec\u7684\u6280\u672f\u53ef\u4ee5\u4e0e Google \u7684\u57fa\u7840\u6a21\u578b\u7ed3\u5408\uff0c\u4e3a\u5176\u63d0\u4f9b\u5feb\u901f\u5b9a\u5236\u5316\u7684\u80fd\u529b\u3002\u540c\u65f6\uff0c\u6211\u4eec\u7684\u6280\u672f\u4e5f\u53ef\u4ee5\u96c6\u6210\u5230 Google \u7684\u5e94\u7528\u5e73\u53f0\u4e2d\uff0c\u4e3a\u5f00\u53d1\u8005\u63d0\u4f9b\u66f4\u4e30\u5bcc\u7684\u5de5\u5177\u9009\u62e9\u3002</p>\n<p>\u7279\u522b\u662f\u5728\u591a\u6a21\u6001\u5e94\u7528\u65b9\u9762\uff0cGoogle \u7684\u6280\u672f\u4f18\u52bf\u660e\u663e\uff0c\u800c\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u53ef\u4ee5\u4e3a\u591a\u6a21\u6001\u5e94\u7528\u63d0\u4f9b\u4e2a\u6027\u5316\u5b9a\u5236\u80fd\u529b\u3002\u4f8b\u5982\uff0c\u5728\u56fe\u50cf\u751f\u6210\u3001\u8bed\u97f3\u5408\u6210\u3001\u89c6\u9891\u5236\u4f5c\u7b49\u5e94\u7528\u4e2d\uff0c\u7528\u6237\u53ef\u4ee5\u901a\u8fc7\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u5feb\u901f\u5b9a\u5236\u7b26\u5408\u81ea\u5df1\u9700\u6c42\u7684 AI \u52a9\u624b\u3002</p>\n<p>\u6211\u4eec\u6b63\u5728\u4e0e Google \u7b49\u56fd\u9645\u5382\u5546\u63a2\u8ba8\u6280\u672f\u5408\u4f5c\u7684\u53ef\u80fd\u6027\uff0c\u5e0c\u671b\u901a\u8fc7\u5f00\u653e\u5408\u4f5c\u5b9e\u73b0\u6280\u672f\u7684\u4e92\u8865\u548c\u5171\u8d62\u3002\u6211\u4eec\u76f8\u4fe1\uff0c\u53ea\u6709\u901a\u8fc7\u5168\u7403\u5408\u4f5c\uff0c\u624d\u80fd\u63a8\u52a8 AI \u6280\u672f\u7684\u5feb\u901f\u53d1\u5c55\u548c\u5e7f\u6cdb\u5e94\u7528\u3002</p>\n<h2>\u9644\u5f55 C\uff1a\u56e2\u961f\u6280\u672f\u5b9e\u529b\u4e0e\u7814\u53d1\u4f53\u7cfb</h2>\n<h3>C.1 \u6838\u5fc3\u6280\u672f\u56e2\u961f\u7684\u5b66\u672f\u80cc\u666f\u4e0e\u5de5\u7a0b\u7ecf\u9a8c</h3>\n<p>\u6211\u4eec\u7684\u6838\u5fc3\u6280\u672f\u56e2\u961f\u6c47\u805a\u4e86\u6765\u81ea\u9876\u5c16\u9662\u6821\u548c\u79d1\u6280\u4f01\u4e1a\u7684\u4f18\u79c0\u4eba\u624d\uff0c\u5728\u4eba\u5de5\u667a\u80fd\u3001\u8ba4\u77e5\u79d1\u5b66\u3001\u8f6f\u4ef6\u5de5\u7a0b\u7b49\u9886\u57df\u5177\u6709\u6df1\u539a\u7684\u5b66\u672f\u80cc\u666f\u548c\u4e30\u5bcc\u7684\u5de5\u7a0b\u7ecf\u9a8c\u3002</p>\n<p>\u56e2\u961f\u7684\u5b66\u672f\u80cc\u666f\u6db5\u76d6\u4e86\u8ba1\u7b97\u673a\u79d1\u5b66\u3001\u8ba4\u77e5\u5fc3\u7406\u5b66\u3001\u8bed\u8a00\u5b66\u3001\u6570\u5b66\u7b49\u591a\u4e2a\u5b66\u79d1\u3002\u8fd9\u79cd\u8de8\u5b66\u79d1\u7684\u5b66\u672f\u80cc\u666f\u4e3a\u6211\u4eec\u7684\u6280\u672f\u521b\u65b0\u63d0\u4f9b\u4e86\u575a\u5b9e\u7684\u7406\u8bba\u57fa\u7840\u3002</p>\n<h3>C.2 \u7814\u53d1\u4f53\u7cfb\u4e0e\u6280\u672f\u521b\u65b0\u673a\u5236</h3>\n<p>\u6211\u4eec\u5efa\u7acb\u4e86\u5b8c\u5584\u7684\u7814\u53d1\u4f53\u7cfb\u548c\u6280\u672f\u521b\u65b0\u673a\u5236\uff0c\u786e\u4fdd\u6280\u672f\u7684\u6301\u7eed\u521b\u65b0\u548c\u5feb\u901f\u8fed\u4ee3\u3002</p>\n<p>\u7814\u53d1\u4f53\u7cfb\u91c7\u7528\u4e86\u654f\u6377\u5f00\u53d1\u7684\u65b9\u6cd5\uff0c\u5c06\u590d\u6742\u7684\u6280\u672f\u9879\u76ee\u5206\u89e3\u4e3a\u591a\u4e2a\u5c0f\u7684\u8fed\u4ee3\u5468\u671f\u3002\u6bcf\u4e2a\u8fed\u4ee3\u5468\u671f\u90fd\u6709\u660e\u786e\u7684\u76ee\u6807\u548c\u53ef\u4ea4\u4ed8\u6210\u679c\uff0c\u786e\u4fdd\u6280\u672f\u53d1\u5c55\u7684\u53ef\u63a7\u6027\u548c\u53ef\u9884\u6d4b\u6027\u3002\u540c\u65f6\uff0c\u6211\u4eec\u5efa\u7acb\u4e86\u5b8c\u5584\u7684\u4ee3\u7801\u7ba1\u7406\u3001\u6d4b\u8bd5\u9a8c\u8bc1\u3001\u90e8\u7f72\u53d1\u5e03\u7b49\u5de5\u7a0b\u6d41\u7a0b\uff0c\u786e\u4fdd\u6280\u672f\u7684\u8d28\u91cf\u548c\u7a33\u5b9a\u6027\u3002</p>\n<p>\u6280\u672f\u521b\u65b0\u673a\u5236\u5305\u62ec\u5b9a\u671f\u7684\u6280\u672f\u5206\u4eab\u3001\u5934\u8111\u98ce\u66b4\u3001\u539f\u578b\u9a8c\u8bc1\u7b49\u6d3b\u52a8\u3002\u6211\u4eec\u9f13\u52b1\u56e2\u961f\u6210\u5458\u63d0\u51fa\u521b\u65b0\u60f3\u6cd5\uff0c\u5e76\u63d0\u4f9b\u5145\u5206\u7684\u8d44\u6e90\u652f\u6301\u8fdb\u884c\u9a8c\u8bc1\u548c\u5b9e\u73b0\u3002\u6211\u4eec\u8fd8\u5efa\u7acb\u4e86\u4e0e\u5916\u90e8\u7814\u7a76\u673a\u6784\u7684\u5408\u4f5c\u5173\u7cfb\uff0c\u901a\u8fc7\u5b66\u672f\u4ea4\u6d41\u548c\u8054\u5408\u7814\u7a76\u63a8\u52a8\u6280\u672f\u521b\u65b0\u3002</p>\n<p>\u6211\u4eec\u7279\u522b\u91cd\u89c6\u6280\u672f\u6587\u6863\u548c\u77e5\u8bc6\u7ba1\u7406\uff0c\u5efa\u7acb\u4e86\u5b8c\u6574\u7684\u6280\u672f\u6587\u6863\u4f53\u7cfb\u548c\u77e5\u8bc6\u5e93\u3002\u6bcf\u4e2a\u6280\u672f\u521b\u65b0\u90fd\u6709\u8be6\u7ec6\u7684\u6587\u6863\u8bb0\u5f55\uff0c\u5305\u62ec\u8bbe\u8ba1\u601d\u8def\u3001\u5b9e\u73b0\u65b9\u6cd5\u3001\u6d4b\u8bd5\u7ed3\u679c\u3001\u5e94\u7528\u573a\u666f\u7b49\u3002\u8fd9\u79cd\u77e5\u8bc6\u7ba1\u7406\u4f53\u7cfb\u4e3a\u6280\u672f\u7684\u4f20\u627f\u548c\u53d1\u5c55\u63d0\u4f9b\u4e86\u91cd\u8981\u4fdd\u969c\u3002</p>\n<h3>C.3 \u6280\u672f\u4e13\u5229\u4e0e\u77e5\u8bc6\u4ea7\u6743\u5e03\u5c40</h3>\n<p>\u6211\u4eec\u4e5f\u79ef\u6781\u53c2\u4e0e\u884c\u4e1a\u6807\u51c6\u7684\u5236\u5b9a\uff0c\u5e0c\u671b\u901a\u8fc7\u6807\u51c6\u5316\u63a8\u52a8\u6280\u672f\u7684\u5e7f\u6cdb\u5e94\u7528\u3002\u6211\u4eec\u5df2\u7ecf\u5411\u76f8\u5173\u6807\u51c6\u5316\u7ec4\u7ec7\u63d0\u4ea4\u4e86\u591a\u9879\u6280\u672f\u63d0\u6848\uff0c\u6d89\u53ca\u8ba4\u77e5\u5fae\u8c03\u7684\u6280\u672f\u89c4\u8303\u3001\u63a5\u53e3\u6807\u51c6\u3001\u5b89\u5168\u8981\u6c42\u7b49\u65b9\u9762\u3002</p>\n<p>\u5728\u5f00\u6e90\u7b56\u7565\u65b9\u9762\uff0c\u6211\u4eec\u91c7\u7528\u4e86\u5e73\u8861\u7684\u77e5\u8bc6\u4ea7\u6743\u7b56\u7565\u3002\u6838\u5fc3\u7684\u6280\u672f\u521b\u65b0\u901a\u8fc7\u4e13\u5229\u4fdd\u62a4\uff0c\u800c\u5e94\u7528\u5c42\u9762\u7684\u6280\u672f\u901a\u8fc7\u5f00\u6e90\u5206\u4eab\u3002\u8fd9\u79cd\u7b56\u7565\u65e2\u4fdd\u62a4\u4e86\u6211\u4eec\u7684\u6280\u672f\u4f18\u52bf\uff0c\u4e5f\u4fc3\u8fdb\u4e86\u6280\u672f\u7684\u666e\u53ca\u5e94\u7528\u3002</p>\n<h2>\u9644\u5f55 D\uff1a\u4ea7\u4e1a\u751f\u6001\u5efa\u8bbe\u4e0e\u6218\u7565\u5408\u4f5c</h2>\n<h3>D.1 \u5f00\u53d1\u8005\u751f\u6001\u7684\u6df1\u5ea6\u5efa\u8bbe</h3>\n<p>\u6211\u4eec\u6df1\u77e5\u5f00\u53d1\u8005\u751f\u6001\u5bf9\u4e8e\u6280\u672f\u6210\u529f\u7684\u91cd\u8981\u6027\uff0c\u56e0\u6b64\u6295\u5165\u4e86\u5927\u91cf\u8d44\u6e90\u8fdb\u884c\u5f00\u53d1\u8005\u751f\u6001\u7684\u5efa\u8bbe\u3002</p>\n<p>\u6211\u4eec\u5efa\u7acb\u4e86\u5b8c\u6574\u7684\u5f00\u53d1\u8005\u652f\u6301\u4f53\u7cfb\uff0c\u5305\u62ec\u6280\u672f\u6587\u6863\u3001\u5f00\u53d1\u5de5\u5177\u3001\u793a\u4f8b\u4ee3\u7801\u3001\u6559\u7a0b\u89c6\u9891\u3001\u5728\u7ebf\u57f9\u8bad\u7b49\u3002\u6280\u672f\u6587\u6863\u91c7\u7528\u4e86\u591a\u5c42\u6b21\u7684\u7ed3\u6784\uff0c\u65e2\u6709\u9762\u5411\u521d\u5b66\u8005\u7684\u5165\u95e8\u6307\u5357\uff0c\u4e5f\u6709\u9762\u5411\u4e13\u5bb6\u7684\u6df1\u5ea6\u6280\u672f\u6587\u6863\u3002\u5f00\u53d1\u5de5\u5177\u5305\u62ec\u8ba4\u77e5\u6a21\u677f\u8bbe\u8ba1\u5668\u3001\u8c03\u8bd5\u5de5\u5177\u3001\u6027\u80fd\u5206\u6790\u5668\u7b49\uff0c\u5927\u5927\u964d\u4f4e\u4e86\u5f00\u53d1\u8005\u7684\u4f7f\u7528\u95e8\u69db\u3002</p>\n<p>\u6211\u4eec\u8fd8\u5efa\u7acb\u4e86\u6d3b\u8dc3\u7684\u5f00\u53d1\u8005\u793e\u533a\uff0c\u5305\u62ec\u6280\u672f\u8bba\u575b\u3001\u5f00\u6e90\u9879\u76ee\u3001\u6280\u672f\u7ade\u8d5b\u7b49\u3002\u5f00\u53d1\u8005\u53ef\u4ee5\u5728\u793e\u533a\u4e2d\u5206\u4eab\u7ecf\u9a8c\u3001\u4ea4\u6d41\u6280\u672f\u3001\u5bfb\u6c42\u5e2e\u52a9\u3002\u6211\u4eec\u5b9a\u671f\u4e3e\u529e\u6280\u672f\u6c99\u9f99\u3001\u5f00\u53d1\u8005\u5927\u4f1a\u7b49\u6d3b\u52a8\uff0c\u4e3a\u5f00\u53d1\u8005\u63d0\u4f9b\u9762\u5bf9\u9762\u4ea4\u6d41\u7684\u673a\u4f1a\u3002</p>\n<h3>D.2 \u4ea7\u4e1a\u5408\u4f5c\u4f19\u4f34\u7f51\u7edc\u7684\u6784\u5efa</h3>\n<p>\u6211\u4eec\u6b63\u5728\u6784\u5efa\u5e7f\u6cdb\u7684\u4ea7\u4e1a\u5408\u4f5c\u4f19\u4f34\u7f51\u7edc\uff0c\u5305\u62ec\u4e91\u670d\u52a1\u63d0\u4f9b\u5546\u3001\u5e94\u7528\u5f00\u53d1\u5546\u3001\u5185\u5bb9\u521b\u4f5c\u5e73\u53f0\u3001\u6559\u80b2\u673a\u6784\u7b49\u3002</p>\n<p>\u4e0e\u4e91\u670d\u52a1\u63d0\u4f9b\u5546\u7684\u5408\u4f5c\u4e3b\u8981\u96c6\u4e2d\u5728\u6280\u672f\u96c6\u6210\u548c\u670d\u52a1\u90e8\u7f72\u65b9\u9762\u3002\u6211\u4eec\u7684\u6280\u672f\u5df2\u7ecf\u4e0e\u591a\u5bb6\u4e3b\u6d41\u4e91\u670d\u52a1\u63d0\u4f9b\u5546\u8fdb\u884c\u4e86\u96c6\u6210\uff0c\u7528\u6237\u53ef\u4ee5\u901a\u8fc7\u4e91\u670d\u52a1\u5e73\u53f0\u76f4\u63a5\u4f7f\u7528\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u3002\u8fd9\u79cd\u5408\u4f5c\u6a21\u5f0f\u5927\u5927\u964d\u4f4e\u4e86\u7528\u6237\u7684\u90e8\u7f72\u6210\u672c\u548c\u6280\u672f\u95e8\u69db\u3002</p>\n<p>\u4e0e\u5e94\u7528\u5f00\u53d1\u5546\u7684\u5408\u4f5c\u4e3b\u8981\u96c6\u4e2d\u5728\u5e94\u7528\u573a\u666f\u7684\u62d3\u5c55\u548c\u89e3\u51b3\u65b9\u6848\u7684\u5f00\u53d1\u65b9\u9762\u3002\u6211\u4eec\u4e0e\u591a\u5bb6\u5e94\u7528\u5f00\u53d1\u5546\u5408\u4f5c\uff0c\u5171\u540c\u5f00\u53d1\u4e86\u9762\u5411\u4e0d\u540c\u884c\u4e1a\u548c\u573a\u666f\u7684\u89e3\u51b3\u65b9\u6848\uff0c\u5305\u62ec\u6559\u80b2\u3001\u5a31\u4e50\u3001\u7535\u5546\u3001\u5ba2\u670d\u7b49\u9886\u57df\u3002</p>\n<p>\u4e0e\u5185\u5bb9\u521b\u4f5c\u5e73\u53f0\u7684\u5408\u4f5c\u4e3b\u8981\u96c6\u4e2d\u5728\u5185\u5bb9\u751f\u6210\u548c\u521b\u4f5c\u8f85\u52a9\u65b9\u9762\u3002\u6211\u4eec\u7684\u6280\u672f\u53ef\u4ee5\u4e3a\u5185\u5bb9\u521b\u4f5c\u8005\u63d0\u4f9b\u4e2a\u6027\u5316\u7684 AI \u52a9\u624b\uff0c\u5e2e\u52a9\u4ed6\u4eec\u63d0\u9ad8\u521b\u4f5c\u6548\u7387\u548c\u8d28\u91cf\u3002</p>\n<p>\u4e0e\u6559\u80b2\u673a\u6784\u7684\u5408\u4f5c\u4e3b\u8981\u96c6\u4e2d\u5728\u4eba\u624d\u57f9\u517b\u548c\u6280\u672f\u63a8\u5e7f\u65b9\u9762\u3002\u6211\u4eec\u4e0e\u591a\u6240\u9ad8\u6821\u5efa\u7acb\u4e86\u5408\u4f5c\u5173\u7cfb\uff0c\u5171\u540c\u5f00\u5c55\u6280\u672f\u7814\u7a76\u548c\u4eba\u624d\u57f9\u517b\u3002\u6211\u4eec\u8fd8\u4e3a\u6559\u80b2\u673a\u6784\u63d0\u4f9b\u6280\u672f\u57f9\u8bad\u548c\u8bfe\u7a0b\u652f\u6301\uff0c\u63a8\u52a8 AI \u6280\u672f\u5728\u6559\u80b2\u9886\u57df\u7684\u5e94\u7528\u3002</p>\n<h3>D.3 \u56fd\u9645\u5316\u53d1\u5c55\u7684\u6218\u7565\u5e03\u5c40</h3>\n<p>\u6211\u4eec\u5236\u5b9a\u4e86\u660e\u786e\u7684\u56fd\u9645\u5316\u53d1\u5c55\u6218\u7565\uff0c\u5e0c\u671b\u5c06\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u63a8\u5e7f\u5230\u5168\u7403\u5e02\u573a\u3002</p>\n<p>\u6280\u672f\u672c\u5730\u5316\u662f\u56fd\u9645\u5316\u53d1\u5c55\u7684\u91cd\u8981\u57fa\u7840\u3002\u6211\u4eec\u6b63\u5728\u9488\u5bf9\u4e0d\u540c\u8bed\u8a00\u3001\u4e0d\u540c\u6587\u5316\u80cc\u666f\u8fdb\u884c\u6280\u672f\u672c\u5730\u5316\uff0c\u5305\u62ec\u8bed\u8a00\u6a21\u578b\u7684\u9002\u914d\u3001\u6587\u5316\u7279\u5f81\u7684\u7406\u89e3\u3001\u4ea4\u4e92\u4e60\u60ef\u7684\u9002\u5e94\u7b49\u3002\u6211\u4eec\u5df2\u7ecf\u5b8c\u6210\u4e86\u82f1\u8bed\u3001\u65e5\u8bed\u3001\u97e9\u8bed\u7b49\u591a\u79cd\u8bed\u8a00\u7684\u6280\u672f\u9002\u914d\uff0c\u672a\u6765\u8fd8\u5c06\u6269\u5c55\u5230\u66f4\u591a\u8bed\u8a00\u3002</p>\n<p>\u5408\u89c4\u6027\u4fdd\u969c\u662f\u56fd\u9645\u5316\u53d1\u5c55\u7684\u91cd\u8981\u524d\u63d0\u3002\u6211\u4eec\u4e25\u683c\u9075\u5b88\u5404\u56fd\u7684\u6cd5\u5f8b\u6cd5\u89c4\u548c\u884c\u4e1a\u6807\u51c6\uff0c\u5305\u62ec\u6570\u636e\u4fdd\u62a4\u3001\u9690\u79c1\u5b89\u5168\u3001\u5185\u5bb9\u5ba1\u6838\u7b49\u65b9\u9762\u7684\u8981\u6c42\u3002\u6211\u4eec\u5efa\u7acb\u4e86\u4e13\u95e8\u7684\u5408\u89c4\u56e2\u961f\uff0c\u8d1f\u8d23\u8ddf\u8e2a\u548c\u5e94\u5bf9\u5404\u56fd\u7684\u6cd5\u89c4\u53d8\u5316\u3002</p>\n<p>\u5e02\u573a\u62d3\u5c55\u662f\u56fd\u9645\u5316\u53d1\u5c55\u7684\u91cd\u8981\u76ee\u6807\u3002\u6211\u4eec\u6b63\u5728\u591a\u4e2a\u56fd\u5bb6\u548c\u5730\u533a\u5efa\u7acb\u672c\u5730\u5316\u7684\u5e02\u573a\u56e2\u961f\uff0c\u8d1f\u8d23\u5e02\u573a\u63a8\u5e7f\u3001\u5ba2\u6237\u670d\u52a1\u3001\u5408\u4f5c\u4f19\u4f34\u53d1\u5c55\u7b49\u5de5\u4f5c\u3002\u6211\u4eec\u8fd8\u53c2\u52a0\u4e86\u591a\u4e2a\u56fd\u9645\u6027\u7684\u6280\u672f\u5c55\u4f1a\u548c\u4f1a\u8bae\uff0c\u63d0\u5347\u6280\u672f\u7684\u56fd\u9645\u77e5\u540d\u5ea6\u3002</p>\n<p>\u6587\u5316\u9002\u5e94\u6027\u662f\u56fd\u9645\u5316\u53d1\u5c55\u7684\u91cd\u8981\u6311\u6218\u3002\u4e0d\u540c\u6587\u5316\u80cc\u666f\u4e0b\u7684\u8ba4\u77e5\u6a21\u5f0f\u548c\u4ea4\u4e92\u4e60\u60ef\u5b58\u5728\u663e\u8457\u5dee\u5f02\uff0c\u6211\u4eec\u9700\u8981\u6df1\u5165\u7814\u7a76\u548c\u7406\u89e3\u8fd9\u4e9b\u5dee\u5f02\uff0c\u786e\u4fdd\u6280\u672f\u80fd\u591f\u63d0\u4f9b\u6587\u5316\u9002\u5e94\u6027\u7684\u670d\u52a1\u3002\u6211\u4eec\u5efa\u7acb\u4e86\u8de8\u6587\u5316\u7814\u7a76\u56e2\u961f\uff0c\u4e13\u95e8\u8d1f\u8d23\u8fd9\u65b9\u9762\u7684\u5de5\u4f5c\u3002</p>\n<h2>\u9644\u5f55 E\uff1a\u6280\u672f\u53d1\u5c55\u7684\u524d\u77bb\u6027\u601d\u8003</h2>\n<h3>E.1 \u4e0b\u4e00\u4ee3\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u6280\u672f\u8def\u7ebf</h3>\n<p>\u6211\u4eec\u5bf9\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u7684\u672a\u6765\u53d1\u5c55\u6709\u7740\u6e05\u6670\u7684\u6280\u672f\u8def\u7ebf\u89c4\u5212\uff0c\u8fd9\u4e9b\u89c4\u5212\u57fa\u4e8e\u6211\u4eec\u5bf9\u6280\u672f\u8d8b\u52bf\u7684\u6df1\u5ea6\u5206\u6790\u548c\u524d\u77bb\u6027\u601d\u8003\u3002</p>\n<p>\u591a\u6a21\u6001\u8ba4\u77e5\u5fae\u8c03\u662f\u6211\u4eec\u7684\u91cd\u8981\u53d1\u5c55\u65b9\u5411\u3002\u5f53\u524d\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4e3b\u8981\u96c6\u4e2d\u5728\u6587\u672c\u9886\u57df\uff0c\u672a\u6765\u6211\u4eec\u5c06\u6269\u5c55\u5230\u56fe\u50cf\u3001\u97f3\u9891\u3001\u89c6\u9891\u7b49\u591a\u6a21\u6001\u9886\u57df\u3002\u591a\u6a21\u6001\u8ba4\u77e5\u5fae\u8c03\u9762\u4e34\u7740\u66f4\u590d\u6742\u7684\u6280\u672f\u6311\u6218\uff0c\u5305\u62ec\u8de8\u6a21\u6001\u7684\u8ba4\u77e5\u4e00\u81f4\u6027\u4fdd\u969c\u3001\u591a\u6a21\u6001\u4fe1\u606f\u7684\u878d\u5408\u5904\u7406\u3001\u591a\u6a21\u6001\u4ea4\u4e92\u7684\u8bbe\u8ba1\u7b49\u3002\u6211\u4eec\u6b63\u5728\u8fdb\u884c\u76f8\u5173\u7684\u6280\u672f\u7814\u7a76\u548c\u539f\u578b\u5f00\u53d1\u3002</p>\n<p>\u81ea\u9002\u5e94\u8ba4\u77e5\u5fae\u8c03\u662f\u6211\u4eec\u7684\u53e6\u4e00\u4e2a\u91cd\u8981\u65b9\u5411\u3002\u5f53\u524d\u7684\u8ba4\u77e5\u5fae\u8c03\u4e3b\u8981\u4f9d\u8d56\u4eba\u5de5\u8bbe\u8ba1\u7684\u8ba4\u77e5\u6a21\u677f\uff0c\u672a\u6765\u6211\u4eec\u5e0c\u671b\u5b9e\u73b0\u81ea\u9002\u5e94\u7684\u8ba4\u77e5\u5fae\u8c03\uff0c\u8ba9\u7cfb\u7edf\u80fd\u591f\u6839\u636e\u7528\u6237\u7684\u4f7f\u7528\u60c5\u51b5\u81ea\u52a8\u8c03\u6574\u548c\u4f18\u5316\u8ba4\u77e5\u7b56\u7565\u3002\u8fd9\u9700\u8981\u7ed3\u5408\u5f3a\u5316\u5b66\u4e60\u3001\u5143\u5b66\u4e60\u3001\u81ea\u76d1\u7763\u5b66\u4e60\u7b49\u5148\u8fdb\u6280\u672f\u3002</p>\n<h3>E.2 \u4e0e\u65b0\u5174\u6280\u672f\u7684\u878d\u5408\u53d1\u5c55</h3>\n<p>\u6211\u4eec\u5bc6\u5207\u5173\u6ce8\u65b0\u5174\u6280\u672f\u7684\u53d1\u5c55\u8d8b\u52bf\uff0c\u79ef\u6781\u63a2\u7d22\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u4e0e\u5176\u4ed6\u65b0\u5174\u6280\u672f\u7684\u878d\u5408\u53d1\u5c55\u3002</p>\n<p>\u4e0e\u8fb9\u7f18\u8ba1\u7b97\u7684\u878d\u5408\u662f\u4e00\u4e2a\u73b0\u5b9e\u7684\u6280\u672f\u9700\u6c42\u3002\u968f\u7740\u7269\u8054\u7f51\u548c\u79fb\u52a8\u8bbe\u5907\u7684\u666e\u53ca\uff0c\u8d8a\u6765\u8d8a\u591a\u7684 AI \u5e94\u7528\u9700\u8981\u5728\u8fb9\u7f18\u8bbe\u5907\u4e0a\u8fd0\u884c\u3002\u6211\u4eec\u6b63\u5728\u7814\u7a76\u8f7b\u91cf\u5316\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\uff0c\u4f7f\u5176\u80fd\u591f\u5728\u8d44\u6e90\u53d7\u9650\u7684\u8fb9\u7f18\u8bbe\u5907\u4e0a\u9ad8\u6548\u8fd0\u884c\u3002</p>\n<p>\u4e0e\u533a\u5757\u94fe\u6280\u672f\u7684\u878d\u5408\u662f\u4e00\u4e2a\u6709\u8da3\u7684\u63a2\u7d22\u65b9\u5411\u3002\u533a\u5757\u94fe\u6280\u672f\u53ef\u4ee5\u4e3a\u8ba4\u77e5\u5fae\u8c03\u63d0\u4f9b\u53bb\u4e2d\u5fc3\u5316\u7684\u4fe1\u4efb\u673a\u5236\uff0c\u4fdd\u62a4\u7528\u6237\u7684\u9690\u79c1\u548c\u6570\u636e\u5b89\u5168\u3002\u6211\u4eec\u6b63\u5728\u7814\u7a76\u57fa\u4e8e\u533a\u5757\u94fe\u7684\u8ba4\u77e5\u5fae\u8c03\u5e73\u53f0\uff0c\u63a2\u7d22\u65b0\u7684\u5546\u4e1a\u6a21\u5f0f\u548c\u6280\u672f\u67b6\u6784\u3002</p>\n<p>\u4e0e\u8111\u673a\u63a5\u53e3\u6280\u672f\u7684\u878d\u5408\u662f\u4e00\u4e2a\u524d\u6cbf\u7684\u7814\u7a76\u65b9\u5411\u3002\u8111\u673a\u63a5\u53e3\u6280\u672f\u53ef\u4ee5\u76f4\u63a5\u83b7\u53d6\u7528\u6237\u7684\u5927\u8111\u4fe1\u53f7\uff0c\u4e3a\u8ba4\u77e5\u5fae\u8c03\u63d0\u4f9b\u66f4\u76f4\u63a5\u3001\u66f4\u51c6\u786e\u7684\u7528\u6237\u53cd\u9988\u3002\u867d\u7136\u8fd9\u4e2a\u65b9\u5411\u8fd8\u5904\u4e8e\u65e9\u671f\u7814\u7a76\u9636\u6bb5\uff0c\u4f46\u6211\u4eec\u76f8\u4fe1\u5b83\u5177\u6709\u5de8\u5927\u7684\u53d1\u5c55\u6f5c\u529b\u3002</p>\n<h3>E.3 \u5bf9 AI \u4ea7\u4e1a\u53d1\u5c55\u7684\u6df1\u5ea6\u601d\u8003</h3>\n<p>\u4f5c\u4e3a AI \u6280\u672f\u7684\u5f00\u53d1\u8005\u548c\u63a8\u52a8\u8005\uff0c\u6211\u4eec\u5bf9 AI \u4ea7\u4e1a\u7684\u53d1\u5c55\u6709\u7740\u6df1\u5ea6\u7684\u601d\u8003\u548c\u72ec\u7279\u7684\u89c1\u89e3\u3002</p>\n<p>\u6211\u4eec\u8ba4\u4e3a AI \u6280\u672f\u7684\u53d1\u5c55\u5c06\u7ecf\u5386\u4ece\u901a\u7528\u5316\u5230\u4e2a\u6027\u5316\u7684\u91cd\u8981\u8f6c\u53d8\u3002\u65e9\u671f\u7684 AI \u6280\u672f\u4e3b\u8981\u5173\u6ce8\u901a\u7528\u80fd\u529b\u7684\u63d0\u5347\uff0c\u5e0c\u671b\u5f00\u53d1\u51fa\u80fd\u591f\u5904\u7406\u5404\u79cd\u4efb\u52a1\u7684\u901a\u7528 AI \u7cfb\u7edf\u3002\u4f46\u968f\u7740\u6280\u672f\u7684\u6210\u719f\u548c\u5e94\u7528\u7684\u6df1\u5165\uff0c\u4e2a\u6027\u5316\u5c06\u6210\u4e3a AI \u6280\u672f\u53d1\u5c55\u7684\u91cd\u8981\u65b9\u5411\u3002\u7528\u6237\u4e0d\u518d\u6ee1\u8db3\u4e8e\u6807\u51c6\u5316\u7684 AI \u670d\u52a1\uff0c\u800c\u662f\u5e0c\u671b\u83b7\u5f97\u7b26\u5408\u81ea\u5df1\u9700\u6c42\u548c\u504f\u597d\u7684\u4e2a\u6027\u5316 AI \u4f53\u9a8c\u3002</p>\n<p>\u6211\u4eec\u8ba4\u4e3a AI \u6280\u672f\u7684\u666e\u53ca\u5c06\u7ecf\u5386\u4ece\u4e13\u4e1a\u5316\u5230\u5e73\u6c11\u5316\u7684\u91cd\u8981\u8fc7\u7a0b\u3002\u65e9\u671f\u7684 AI \u6280\u672f\u4e3b\u8981\u670d\u52a1\u4e8e\u4e13\u4e1a\u7528\u6237\u548c\u5927\u578b\u4f01\u4e1a\uff0c\u666e\u901a\u7528\u6237\u5f88\u96be\u63a5\u89e6\u548c\u4f7f\u7528\u3002\u4f46\u968f\u7740\u6280\u672f\u95e8\u69db\u7684\u964d\u4f4e\u548c\u5de5\u5177\u7684\u5b8c\u5584\uff0cAI \u6280\u672f\u5c06\u9010\u6e10\u666e\u53ca\u5230\u666e\u901a\u7528\u6237\u3002\u6211\u4eec\u7684\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u6b63\u662f\u8fd9\u79cd\u5e73\u6c11\u5316\u8d8b\u52bf\u7684\u91cd\u8981\u4f53\u73b0\u3002</p>\n<p>\u6211\u4eec\u8ba4\u4e3a AI \u4ea7\u4e1a\u7684\u53d1\u5c55\u5c06\u7ecf\u5386\u4ece\u7ade\u4e89\u5230\u5408\u4f5c\u7684\u91cd\u8981\u8f6c\u53d8\u3002\u65e9\u671f\u7684 AI \u4ea7\u4e1a\u4e3b\u8981\u662f\u5404\u5bb6\u4f01\u4e1a\u7684\u72ec\u7acb\u7ade\u4e89\uff0c\u6bcf\u5bb6\u4f01\u4e1a\u90fd\u5e0c\u671b\u5efa\u7acb\u81ea\u5df1\u7684\u6280\u672f\u58c1\u5792\u548c\u751f\u6001\u4f53\u7cfb\u3002\u4f46\u968f\u7740\u6280\u672f\u7684\u590d\u6742\u5316\u548c\u5e94\u7528\u7684\u591a\u6837\u5316\uff0c\u5355\u4e00\u4f01\u4e1a\u5f88\u96be\u8986\u76d6\u6240\u6709\u7684\u6280\u672f\u9886\u57df\u548c\u5e94\u7528\u573a\u666f\u3002\u672a\u6765\u7684 AI \u4ea7\u4e1a\u5c06\u66f4\u591a\u5730\u4f9d\u8d56\u5408\u4f5c\u548c\u751f\u6001\u5efa\u8bbe\uff0c\u901a\u8fc7\u5f00\u653e\u5408\u4f5c\u5b9e\u73b0\u5171\u8d62\u53d1\u5c55\u3002</p>\n<p>\u6211\u4eec\u8ba4\u4e3a AI \u6280\u672f\u7684\u53d1\u5c55\u5fc5\u987b\u59cb\u7ec8\u575a\u6301\u4ee5\u4eba\u4e3a\u672c\u7684\u4ef7\u503c\u7406\u5ff5\u3002\u6280\u672f\u7684\u6700\u7ec8\u76ee\u7684\u662f\u4e3a\u4eba\u7c7b\u670d\u52a1\uff0c\u63d0\u5347\u4eba\u7c7b\u7684\u751f\u6d3b\u8d28\u91cf\u548c\u5de5\u4f5c\u6548\u7387\u3002\u6211\u4eec\u5728\u6280\u672f\u5f00\u53d1\u8fc7\u7a0b\u4e2d\u59cb\u7ec8\u575a\u6301\u8fd9\u4e00\u7406\u5ff5\uff0c\u786e\u4fdd\u6280\u672f\u7684\u53d1\u5c55\u7b26\u5408\u4eba\u7c7b\u7684\u957f\u8fdc\u5229\u76ca\u3002\u6211\u4eec\u4e5f\u547c\u5401\u6574\u4e2a\u884c\u4e1a\u5171\u540c\u575a\u6301\u8fd9\u4e00\u7406\u5ff5\uff0c\u63a8\u52a8 AI \u6280\u672f\u7684\u8d1f\u8d23\u4efb\u53d1\u5c55\u3002</p>\n<hr/>\n<p><em>\u672c\u6280\u672f\u767d\u76ae\u4e66\u4ee3\u8868\u4e86\u5e7b\u5b99\u667a\u80fd\u56e2\u961f\u5728\u8ba4\u77e5\u5fae\u8c03\u6280\u672f\u9886\u57df\u7684\u6700\u65b0\u7814\u7a76\u6210\u679c\u548c\u6280\u672f\u79ef\u7d2f\u3002\u6211\u4eec\u5c06\u7ee7\u7eed\u81f4\u529b\u4e8e\u6280\u672f\u521b\u65b0\u548c\u5f00\u653e\u5408\u4f5c\uff0c\u63a8\u52a8 AI \u6280\u672f\u7684\u53d1\u5c55\u548c\u5e94\u7528\uff0c\u4e3a\u4eba\u7c7b\u793e\u4f1a\u521b\u9020\u66f4\u5927\u7684\u4ef7\u503c\u3002</em></p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/koala9527", 
        "name": "koala9527", 
        "avatar": "https://cdn.v2ex.com/avatar/53d3/8cb7/419379_large.png?m=1770049432"
      }, 
      "url": "https://www.v2ex.com/t/945424", 
      "date_modified": "2023-06-03T02:46:41+00:00", 
      "content_html": "<p>\u6392\u67e5\u4e86\u597d\u534a\u5929\n\u6253\u4e86\u5341\u51e0\u4e2a\u955c\u50cf\u6d4b\u8bd5\u95ee\u9898\uff0c\u5bb9\u5668\u4e00\u8fd0\u884c\u5c31\u542f\u52a8\u5931\u8d25\uff0c\u5728\u4e91\u670d\u52a1\u5668\uff0c\u81ea\u5df1\u7684 Win10 \u7535\u8111\u5b8c\u7f8e\u8fd0\u884c\uff0c\n\u5bb9\u5668\u7684\u72b6\u6001\u662f\uff1a</p>\n<pre><code>Restarting (132) 27 seconds ago\n</code></pre>\n<p>docker log \u4e5f\u6ca1\u6709\u770b\u5230\u5bb9\u5668\u65e5\u5fd7</p>\n<p>\u641c\u7d22\u4e86\u4e00\u4e0b\u8fd9\u4e2a\u62a5\u9519\u53d1\u73b0\u662f AVX \u6307\u4ee4\u96c6\u7684\u95ee\u9898\n\u5b98\u65b9\u6587\u6863\u8bf4\u660e\uff1a\u4ece TensorFlow 1.6 \u5f00\u59cb\uff0c\u4e8c\u8fdb\u5236\u6587\u4ef6\u4f7f\u7528 AVX \u6307\u4ee4\uff0c\u8fd9\u4e9b\u6307\u4ee4\u53ef\u80fd\u65e0\u6cd5\u5728\u65e7\u7248 CPU \u4e0a\u8fd0\u884c\u3002</p>\n<p>\u770b\u4e86\u4e00\u4e9b\u8fd9\u4e2a\u673a\u5668\u7684\u662f\u5426\u5305\u542b AVX \u6307\u4ee4\u96c6</p>\n<pre><code>sudo cat /proc/cpuinfo | grep avx\n</code></pre>\n<p>\u7ed3\u679c\u771f\u7684\u6ca1\u6709</p>\n<p>\u7b2c\u4e00\u9047\u5230\u8fd9\u4e2a\u5751\uff0c\u5206\u4eab\u4e00\u4e0b\uff0c\u95ee\u4e0b V \u53cb\u4eec\u662f\u5426\u662f\u65e0\u89e3\u7684\u95ee\u9898\u5440\uff1f</p>\n", 
      "date_published": "2023-06-03T02:39:03+00:00", 
      "title": "\u60f3\u5728 N5015 \u5c0f\u4e3b\u673a\u4e0a\u90e8\u7f72\u4e00\u4e2a Tensorflow \u7684\u9879\u76ee\uff0c\u6ca1\u60f3\u5230\u8fd9\u4e2a CPU \u4e0d\u652f\u6301 AVX \u6307\u4ee4\u96c6", 
      "id": "https://www.v2ex.com/t/945424"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/lixyz", 
        "name": "lixyz", 
        "avatar": "https://cdn.v2ex.com/avatar/f8d9/e12f/88942_large.png?m=1419823074"
      }, 
      "url": "https://www.v2ex.com/t/923261", 
      "title": "\u6709\u6ca1\u6709\u641e\u56fe\u5f62\u7b97\u6cd5\u7684\u5144\u5f1f\uff0c\u6709\u4e2a\u6d3b\u513f", 
      "id": "https://www.v2ex.com/t/923261", 
      "date_published": "2023-03-11T15:44:05+00:00", 
      "content_html": "<p>\u9700\u6c42\uff1a</p>\n<p>\u7ed9\u5b9a\u4e00\u4e2a\u4eba\u7684\u9762\u90e8\u7167\u7247\uff0c\u53ef\u4ee5\u51c6\u786e\u7684\u5206\u6790\u51fa\u8fd9\u4e2a\u4eba\u7269\u9762\u90e8\u7279\u5f81\u7684\u989c\u8272\uff0c\u5305\u542b\u4f46\u4e0d\u9650\u4e8e\uff1a\u4e94\u5b98\u3001\u5934\u53d1...</p>\n<p>\u6700\u597d\u4eba\u5728\u5317\u4eac</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/fenglala", 
        "name": "fenglala", 
        "avatar": "https://cdn.v2ex.com/avatar/4342/5bdd/413366_large.png?m=1726298373"
      }, 
      "url": "https://www.v2ex.com/t/871507", 
      "title": "tensorflow 20 \u79cd\u8bed\u8a00\u6587\u6863\u53ea\u6709\u7b80\u4e2d\u6587\u6863\u8fc7\u65f6\uff0c\u4e5f\u4e0d\u63d0\u793a\u7528\u6237\u8fc7\u65f6", 
      "id": "https://www.v2ex.com/t/871507", 
      "date_published": "2022-08-08T12:27:56+00:00", 
      "content_html": "<p><a href=\"https://www.tensorflow.org/lite/performance/gpu_advanced?hl=zh-cn#android_cc\" rel=\"nofollow\">https://www.tensorflow.org/lite/performance/gpu_advanced?hl=zh-cn#android_cc</a></p>\n<pre><code class=\"language-sh\">bazel build -c opt --config android_arm64 tensorflow/lite/delegates/gpu:gl_delegate                  # for static library\nbazel build -c opt --config android_arm64 tensorflow/lite/delegates/gpu:libtensorflowlite_gpu_gl.so  # for dynamic library\n</code></pre>\n<p><a href=\"https://www.tensorflow.org/lite/performance/gpu_advanced?hl=en-us#android_cc\" rel=\"nofollow\">https://www.tensorflow.org/lite/performance/gpu_advanced?hl=en-us#android_cc</a></p>\n<pre><code class=\"language-sh\">bazel build -c opt --config android_arm64 tensorflow/lite/delegates/gpu:delegate                           # for static library\nbazel build -c opt --config android_arm64 tensorflow/lite/delegates/gpu:libtensorflowlite_gpu_delegate.so  # for dynamic library\n</code></pre>\n<p>\u4e0d\u8bf4\u767d\u767d\u82b1\u65f6\u95f4\u6392\u9519\uff0c\u8fde\u4e2a\u8be5\u6587\u6863\u5df2\u8fc7\u65f6\u7684\u63d0\u793a\u90fd\u6ca1\u6709\uff0c\u7740\u5b9e\u8ba9\u4eba\u4e0d\u9002\u3002</p>\n<p>\u8fc7\u4e86\u4e00\u4e0b 20 \u79cd\u8bed\u8a00\u7684\u6587\u6863\uff0c\u5176\u4ed6\u6240\u6709\u8bed\u8a00\u90fd\u662f\u65b0\u7684<code>gpu:delegate</code>\uff0c\u552f\u72ec\u7b80\u4e2d\u6587\u6863\u662f\u65e7\u7684<code>gpu:gl_delegate</code>\u3002\u4e0d\u6e05\u695a TensorFlow \u56e2\u961f\u662f\u4e0d\u662f\u4f1a\u6c49\u8bed\u7684\u5c11\uff0c\u4f46\u662f\u7ed9\u4e2a\u6807\u8bc6\u63d0\u793a\u6211\u6587\u6863\u662f\u65e7\u7684\u5e94\u8be5\u4e0d\u591a\u5427\u3002</p>\n<p>\u63d0\u4e86\u4e00\u4e2a\u6587\u6863 issue\n<a href=\"https://github.com/tensorflow/tensorflow/issues/57039\" rel=\"nofollow\">https://github.com/tensorflow/tensorflow/issues/57039</a></p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/MuSik", 
        "name": "MuSik", 
        "avatar": "https://cdn.v2ex.com/avatar/0dc2/33c4/490738_large.png?m=1718607831"
      }, 
      "url": "https://www.v2ex.com/t/855320", 
      "date_modified": "2022-05-25T14:47:16+00:00", 
      "content_html": "<p>\u4f7f\u7528\u9879\u76ee\u7528\u5230\u4e86 PixelLib \u56fe\u50cf\u5206\u5272\u8fd9\u4e2a\u5e93<br/>\n\u4f46\u662f\u5bfc\u5165\u4f7f\u7528\u7684\u65f6\u5019\u62a5\u8fd9\u4e2a\u9519\u3002\n<img alt=\"\" class=\"embedded_image\" loading=\"lazy\" referrerpolicy=\"no-referrer\" rel=\"noreferrer\" src=\"https://s3.bmp.ovh/imgs/2022/05/25/d6a01b6839a5133b.png\"/>\n\u8bf7\u95ee\u5404\u4f4d\u5927\u4f6c\uff0c\u8fd9\u600e\u4e48\u89e3\u51b3\uff1f\u975e\u5e38\u611f\u8c22\uff01</p>\n", 
      "date_published": "2022-05-25T14:44:09+00:00", 
      "title": "\u5173\u4e8e tensorflow \u5f15\u7528 PixelLib \u56fe\u50cf\u5206\u5272\u5e93\u7684\u95ee\u9898", 
      "id": "https://www.v2ex.com/t/855320"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/wonderwall2237", 
        "name": "wonderwall2237", 
        "avatar": "https://cdn.v2ex.com/gravatar/b5b1f428c40b4e75eb603117fa42dce1?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/848263", 
      "title": "\u4f60\u5011\u662f\u53bb\u54ea\u627e\u6578\u64da \u53bb\u8a13\u7df4 deep learning tensorflow \u7684?", 
      "id": "https://www.v2ex.com/t/848263", 
      "date_published": "2022-04-20T15:42:56+00:00", 
      "content_html": "\u597d\u5947\u4f60\u5011\u600e\u6a23\u53ef\u4ee5\u627e\u5230\u9019\u9ebc\u591a\u8cc7\u6599\u548c\u76f8\u61c9\u7684\u7b54\u6848\u53bb\u8a13\u7df4 deep learning?"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/leven87", 
        "name": "leven87", 
        "avatar": "https://cdn.v2ex.com/gravatar/b4497986025202f2280dc8497ab80cb7?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/799699", 
      "title": "\u6709\u6ca1\u6709\u4e00\u4e2a\u5408\u9002\u7684\u5f00\u6e90\u7ba1\u7406\u7cfb\u7edf\uff0c\u53ef\u4ee5\u7ba1\u7406 tensorflow \u7684 \u591a\u673a\u591a\u5361\u7684\u8bad\u7ec3", 
      "id": "https://www.v2ex.com/t/799699", 
      "date_published": "2021-09-03T07:55:11+00:00", 
      "content_html": "\u8bf7\u6559\uff0c\u6709\u6ca1\u6709\u4e00\u4e2a\u5408\u9002\u7684\u5f00\u6e90\u7ba1\u7406\u7cfb\u7edf\uff0c\u53ef\u4ee5\u7ba1\u7406 tensorflow docker \u7684 \u591a\u673a\u591a\u5361\u7684\u8bad\u7ec3\u3002\u4e3b\u8981\u529f\u80fd\u662f\uff1a<br />1. \u8bbe\u7f6e\u8fd0\u884c\u7684\u8282\u70b9\u6570\u76ee(docker \u6570\u76ee\uff09;<br />2. \u542f\u52a8\u8fd0\u884c\u8bad\u7ec3\u811a\u672c\uff1b<br />3. \u89c2\u5bdf\u548c\u76d1\u63a7\u8bad\u7ec3\u8fdb\u7a0b\u7b49\u3002<br /><br />\u8c37\u6b4c\u4e86\u4e00\u4e0b\uff0c\u6ca1\u6709\u4ec0\u4e48\u601d\u8def\uff0c\u8c22\u8c22\uff01"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/waitfor", 
        "name": "waitfor", 
        "avatar": "https://cdn.v2ex.com/avatar/4ea4/7504/525756_large.png?m=1612430322"
      }, 
      "url": "https://www.v2ex.com/t/751322", 
      "title": "Tensorflow 1.x \u7248\u672c\u5982\u4f55\u5bf9 Tensorflow hub \u7684\u6a21\u578b\u8fdb\u884c\u5fae\u8c03\uff1f", 
      "id": "https://www.v2ex.com/t/751322", 
      "date_published": "2021-02-04T09:25:25+00:00", 
      "content_html": "<p>\u6211\u770b\u4e86\u5b98\u65b9\u7684\u6587\u6863\uff0c\u5c31\u51e0\u884c\u4ee3\u7801\uff0c\u6709\u6ca1\u6709\u4e00\u4e2a\u5168\u7684\u4ee3\u7801\u5440\u3002\u3002\n\u5b98\u65b9\u7684\u4ee3\u7801\u5168\u662f TF2 keras \u7684\uff0c\u96be\u53d7\uff01\n\u6025\u6025\u6025\uff01\n<code><a href=\"https://hub.tensorflow.google.cn/google/imagenet/resnet_v2_50/classification/3\" rel=\"nofollow\">https://hub.tensorflow.google.cn/google/imagenet/resnet_v2_50/classification/3</a></code>\u4ee5\u8fd9\u4e2a\u4e3a\u4f8b\uff0c\u5c31\u53ea\u6709\u51e0\u884c\u4ee3\u7801\uff0c</p>\n<pre><code class=\"language-python\">module = hub.Module(\"https://hub.tensorflow.google.cn/google/imagenet/resnet_v2_50/classification/3\",\n                    trainable=True, tags={\"train\"})\nlogits = module(inputs=dict(images=images, batch_norm_momentum=0.997),\n                signature=\"image_classification_with_bn_hparams\")\n</code></pre>\n<p>\u5230\u5e95\u5982\u4f55\u5fae\u8c03?</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/alabulei", 
        "name": "alabulei", 
        "avatar": "https://cdn.v2ex.com/gravatar/c47ba7e0cf7c2e2b6408ed16afae84ea?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/751194", 
      "title": "\u7528 Serverless \u67b6\u6784\u63d0\u4f9b TensorFlow AI \u63a8\u7406\u51fd\u6570", 
      "id": "https://www.v2ex.com/t/751194", 
      "date_published": "2021-02-04T03:36:43+00:00", 
      "content_html": "<p>\u4e3b\u8981\u6280\u672f\u662f Rust \u51fd\u6570\u3001WebAssembly \u865a\u62df\u673a\u3001Tencent serverless\uff0c\u63d0\u4f9b\u4e86\u4e00\u4e2a\u6a21\u677f\uff0c\u8ba9\u5f00\u53d1\u8005\u65b9\u4fbf\u5feb\u901f\u5730\u5c06 AI \u6a21\u578b\u7528\u5728 web \u5e94\u7528\u4e2d\u3002\n\u8be6\u60c5\u70b9\u51fb<a href=\"https://mp.weixin.qq.com/s?__biz=MzI2MjkxNjA2Mg==&amp;mid=2247483919&amp;idx=1&amp;sn=8b35d8826b6c4d72beed73b59a76c028&amp;chksm=ea4290bedd3519a8ef509ac67b4ccd6985f7058e43f148f94eeceff71958b573222b8513b79d&amp;token=149204981&amp;lang=zh_CN#rd\" rel=\"nofollow\">\u8fd9\u91cc</a></p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/pineapplex", 
        "name": "pineapplex", 
        "avatar": "https://cdn.v2ex.com/gravatar/9b195b1c2cdfde138a6b49b878efe172?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/744688", 
      "title": "\u8bf7\u6559\u4e3a\u4f55\u5728\u865a\u62df\u673a VirualBox \u4e2d\u8fd0\u884c\u7684 Ubuntu20\uff0c\u6ca1\u6cd5\u52a0\u8f7d tensorflow\u3002", 
      "id": "https://www.v2ex.com/t/744688", 
      "date_published": "2021-01-13T12:23:09+00:00", 
      "content_html": "\u6211\u7684 tensorflow \u5728 WIN10 \u7684\u7cfb\u7edf\u4e2d\u5df2\u7ecf\u80fd\u591f\u6b63\u5e38\u8fd0\u884c\uff0c\u5e76\u4e14\u88c5\u4e86 cuda \u548c cudnn\uff0c\u80fd\u591f\u6b63\u5e38\u4f7f\u7528 GPU \u4e86\u3002<br />\u6211\u7a81\u53d1\u5947\u60f3\uff0c\u60f3\u5728\u81ea\u5df1\u7684\u865a\u62df\u673a Ubuntu \u4e0a\u4e5f\u5f00\u4e2a tf\uff0c\u8bf4\u4e0d\u5b9a\u53ef\u4ee5\u5c1d\u8bd5\u4e00\u4e0b\u5206\u5e03\u5f0f\uff1f\u7ed3\u679c\u662f\u597d\u4e0d\u5bb9\u6613\u5b89\u4e0a tensorflow \u540e\uff0c\u65e0\u6cd5\u8fd0\u884c\u3002<br />\u5f53 import tensorflow \u540e\uff0c\u76f4\u63a5\u63d0\u793a<br />F tensorflow/core/platform/<a target=\"_blank\" href=\"http://cpu_feature_guard.cc:38\" rel=\"nofollow noopener\">cpu_feature_guard.cc:38</a>] The TensorFlow library was compiled to use FMA instructions, but these aren't available on your machine.<br />\u4e0d\u77e5\u9053\u95ee\u9898\u5728\u54ea\u91cc\u2026\u2026"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/herozzm", 
        "name": "herozzm", 
        "avatar": "https://cdn.v2ex.com/avatar/620b/1efe/69595_large.png?m=1694144243"
      }, 
      "url": "https://www.v2ex.com/t/704530", 
      "date_modified": "2020-09-05T12:23:49+00:00", 
      "content_html": "<p>\u7528\u4e86 pyton\uff0c\u8fd0\u884c\u4e00\u6bb5\u65f6\u95f4\u540e\uff0c\u5185\u5b58\u5360\u7528\u5230\u4e86\u51e0\u5341 g\uff0c\u6700\u540e\u53ea\u80fd ctrl+c \u5173\u95ed python \u811a\u672c\u91cd\u65b0\u6267\u884c\uff0c\u53c8\u4ece\u5934\u7d2f\u8ba1\uff0c\u6700\u540e\u7206\u5185\u5b58\u4e86\uff0c\u597d\u7d2f\uff0c\u8fd9\u6837\u6709\u89e3\u5417\uff1f</p>\n<pre><code class=\"language-python\">\n   start:\n   \n   # \u67e5\u8be2\u9700\u8981\u8ba1\u7b97\u7684\u6570\u636e lists\n   sql exec...\n\n   # \u5faa\u73af\u8ba1\u7b97\n   for list in lists {\n       tf \u8ba1\u7b97 list \u547d\u540d\u5b9e\u4f53...\n   }    \n   \n\n    sleep(1)\n    \n    goto start\n</code></pre>\n", 
      "date_published": "2020-09-05T12:14:46+00:00", 
      "title": "\u6211\u7528\u4e86 tensorflow \u5faa\u73af\u4ece\u6570\u636e\u4e2d\u63d0\u53d6\u547d\u540d\u5b9e\u4f53\uff0c\u5185\u5b58\u4e00\u76f4\u7d2f\u8ba1\u4e0d\u91ca\u653e", 
      "id": "https://www.v2ex.com/t/704530"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/leven87", 
        "name": "leven87", 
        "avatar": "https://cdn.v2ex.com/gravatar/b4497986025202f2280dc8497ab80cb7?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/698598", 
      "date_modified": "2020-08-15T17:34:42+00:00", 
      "content_html": "\u6211\u5728\u4e91\u670d\u52a1\u5668\u4e0a \uff08 1G, 1cpu \uff09, \u603b\u662f\u63d0\u793a\u5185\u5b58\u4e0d\u591f\u3002<br />\u8bf7\u95ee tensorflow2 \u5b89\u88c5\u6700\u4f4e\u7684\u5185\u5b58\u8981\u6c42\uff0c\u6216\u8005\u6709\u4ec0\u4e48\u65b9\u6cd5\u53ef\u4ee5\u907f\u514d\u8fd9\u4e2a\u63d0\u793a\uff1f \u8c22\u8c22\u5927\u5bb6~", 
      "date_published": "2020-08-15T17:33:31+00:00", 
      "title": "\u8bf7\u95ee tensorflow2 \u5b89\u88c5\u6700\u4f4e\u7684\u5185\u5b58\u8981\u6c42\uff1f", 
      "id": "https://www.v2ex.com/t/698598"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/herozzm", 
        "name": "herozzm", 
        "avatar": "https://cdn.v2ex.com/avatar/620b/1efe/69595_large.png?m=1694144243"
      }, 
      "url": "https://www.v2ex.com/t/680580", 
      "title": "\u60f3\u81ea\u5df1\u8bad\u7ec3 tensorflow \u8bc6\u522b\u67d0\u4e2a\u7f51\u7ad9\u9a8c\u8bc1\u7801\u7684\u6a21\u578b\uff0c\u8c01\u80fd\u53d1\u4e00\u4e2a\u8a00\u7b80\u610f\u8d45\u7684\u6559\u7a0b\uff0c\u4f1a Python", 
      "id": "https://www.v2ex.com/t/680580", 
      "date_published": "2020-06-11T02:58:26+00:00", 
      "content_html": "<p>\u8c22\u8c22\u8001\u94c1</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/seers", 
        "name": "seers", 
        "avatar": "https://cdn.v2ex.com/avatar/46ae/84c9/73974_large.png?m=1773071488"
      }, 
      "url": "https://www.v2ex.com/t/677684", 
      "date_modified": "2020-07-14T17:27:13+00:00", 
      "content_html": "\u6bd4\u8f83\u57fa\u7840\u7684\u6700\u597d\uff0c\u5bf9\u7740\u5b98\u65b9 doc \u8dd1\u4e86\u4e00\u4e0b demo\uff0c\u7a0d\u5fae\u719f\u6089\u4e86\u4e0b\uff0c\u4f46\u662f\u5f88\u591a\u540d\u8bcd\u4e0d\u662f\u5f88\u61c2", 
      "date_published": "2020-06-02T00:41:43+00:00", 
      "title": "Tensorflow \u6709\u6ca1\u6709\u5165\u95e8\u4e66\u7c4d\u63a8\u8350", 
      "id": "https://www.v2ex.com/t/677684"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/herozzm", 
        "name": "herozzm", 
        "avatar": "https://cdn.v2ex.com/avatar/620b/1efe/69595_large.png?m=1694144243"
      }, 
      "url": "https://www.v2ex.com/t/673051", 
      "date_modified": "2020-05-18T16:01:01+00:00", 
      "content_html": "<p>\u672c\u6765\u8981\u5b89\u88c5 gpu \u7248\u672c\uff0c\u5b89\u88c5\u663e\u5361\u9a71\u52a8\u5e76\u5b89\u88c5\u548c nvidia \u5bb9\u5668\u5de5\u5177\u5305\u540e\uff0c\u8fd0\u884c<code>docker run --gpus all --rm nvidia/cuda nvidia-smi</code>\uff0c\u53ef\u4ee5\u770b\u5230\u6b63\u5e38\u7ed3\u679c\uff0c\u7136\u540e\u6b63\u5e38\u5b89\u88c5 tensorflow-gpu docker \u5bb9\u5668\uff08\u91cc\u9762\u7684 tengsorflow \u53ef\u4ee5\u6b63\u5e38\u8fd0\u884c\uff09\uff0c\u5b8c\u6210\u540e exit \u9000\u51fa ssh\uff0c\u8fc7\u4e00\u6bb5\u65f6\u95f4\u518d\u6b21\u94fe\u63a5\u518d\u6b21\u8fd0\u884c\u53d1\u73b0<code>docker run --gpus all --rm nvidia/cuda nvidia-smi</code>\u4f1a\u5361\u6b7b\uff0c\u540c\u65f6 tensorflow \u5bb9\u5668\u91cc\u9762\u63d0\u793a\u65e0\u6cd5\u627e\u5230\u663e\u5361\uff0c\u800c\u8fd0\u884c\u5176\u4ed6\u975e docker \u7684 gpus \u547d\u4ee4\u6b63\u5e38\uff0c\u6709\u4eba\u9047\u5230\u5417\uff1f\u767e\u601d\u4e0d\u5f97\u5176\u89e3\uff0c\u6211\u5c1d\u8bd5\u91cd\u65b0\u91cd\u542f\u5e76\u4ece\u5b89\u88c5\u663e\u5361\u9a71\u52a8\u5f00\u59cb\u53c8\u6062\u590d\u6b63\u5e38\uff0c\u7136\u540e\u8fc7\u6bb5\u65f6\u95f4\u540e\u53c8\u5931\u6548</p>\n", 
      "date_published": "2020-05-18T15:58:07+00:00", 
      "title": "\u5728 debian \u4e2d\u7528 docker \u7248\u672c\u7684 tensorflow-gpu \u5931\u6548\u95ee\u9898", 
      "id": "https://www.v2ex.com/t/673051"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/honeyshine75", 
        "name": "honeyshine75", 
        "avatar": "https://cdn.v2ex.com/gravatar/210da5b71b8f0ed37564405ba057d57b?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/665360", 
      "title": "\u56fe\u50cf\u98ce\u683c\u8fc1\u79fb\u95ee\u9898\uff0c\u6c42\u52a9", 
      "id": "https://www.v2ex.com/t/665360", 
      "date_published": "2020-04-23T07:47:09+00:00", 
      "content_html": "<p>\u6c42\u52a9\u5404\u4f4d\u5927\u4f6c\uff0c\u5e2e\u5fd9\u770b\u770b\u4ee3\u7801\uff08\u5f88\u597d\u770b\u7684\uff0c\u6211 u \u5199\u5f88\u591a\u6ce8\u91ca\uff0c\u5e2e\u5e2e\u65b0\u624b\u5427\uff09\n\u4ee3\u7801\u5982\u4e0b\uff1a <a href=\"https://colab.research.google.com/drive/1tONEeHZB3p9lzLivV-BMxR1p_8Qm6Uoo\" rel=\"nofollow\">https://colab.research.google.com/drive/1tONEeHZB3p9lzLivV-BMxR1p_8Qm6Uoo</a></p>\n<p>\u8fd9\u51e0\u5929\u5728\u770b\u673a\u5668\u89c6\u89c9\uff0c\u770b\u5230\u4e86\u56fe\u7247\u98ce\u683c\u8fc1\u79fb\uff0c\u672c\u6253\u7b97\u81ea\u5df1\u7528 tensorflow \u6539\u5199\u5e76\u8bad\u7ec3\u4e00\u4e2a\uff0c\u4f46\u662f\u5982\u4eca\u9047\u5230\u4e86\u56f0\u96be\uff1a</p>\n<ol>\n<li>\n<p>\u53c2\u8003\u8fd9\u4e2a\u4ee3\u7801\uff1a <a href=\"https://github.com/CortexFoundation/StyleTransferTrilogy/tree/master/keras_version\" rel=\"nofollow\">https://github.com/CortexFoundation/StyleTransferTrilogy/tree/master/keras_version</a></p>\n</li>\n<li>\n<p>\u4fee\u6539\u4e86\u4e0a\u9762\u4ee3\u7801\u4e2d\u7684 style_loss \u8ba1\u7b97\u65b9\u6cd5\u548c content \u3001style \u5c42\uff0c\u6309\u7167 tensorflow \u7684\u6559\u7a0b\u6539\u7684 <a href=\"https://www.tensorflow.org/tutorials/generative/style_transfer#%E5%AE%9A%E4%B9%89%E5%86%85%E5%AE%B9%E5%92%8C%E9%A3%8E%E6%A0%BC%E7%9A%84%E8%A1%A8%E7%A4%BA\" rel=\"nofollow\">https://www.tensorflow.org/tutorials/generative/style_transfer#%E5%AE%9A%E4%B9%89%E5%86%85%E5%AE%B9%E5%92%8C%E9%A3%8E%E6%A0%BC%E7%9A%84%E8%A1%A8%E7%A4%BA</a></p>\n</li>\n<li>\n<p>\u4e3b\u8981\u95ee\u9898\uff1a\u65e0\u6cd5\u6536\u655b\uff0c\u6bcf\u6b21\u8bad\u7ec3\u5230\u540e\u671f\u90fd\u4f1a\u53d8\u6210\u4e00\u4e2a\u57fa\u672c\u56fa\u5b9a\u7684\u6a21\u5f0f\uff0c\u6362\u4e86 style \u56fe\u7247\u540e\u751f\u6210\u7684\u56fe\u7247\u6539\u52a8\u5927\u4e0d\u3002</p>\n</li>\n</ol>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/herozzm", 
        "name": "herozzm", 
        "avatar": "https://cdn.v2ex.com/avatar/620b/1efe/69595_large.png?m=1694144243"
      }, 
      "url": "https://www.v2ex.com/t/660798", 
      "date_modified": "2020-04-09T06:42:12+00:00", 
      "content_html": "<p>\u5df2\u7ecf\u662f\u8bad\u7ec3\u597d\u7684\u6a21\u578b\u4e86\uff0c\u6295\u5165\u751f\u4ea7\u73af\u5883\u4e5f\u5f88\u5403\u7b97\u529b\uff0c\u4e00\u822c\u7684\u670d\u52a1\u5668\u6ca1\u6709\u72ec\u7acb\u663e\u5361\u554a\uff0c\u96be\u9053\u53ea\u80fd\u8dd1 cpu \u7248\u672c,cpu \u901f\u5ea6\u5f88\u6162\uff0c\u6ca1\u6cd5\u7528\u554a\uff0c\u6216\u8005\u7ed9\u670d\u52a1\u5668\u641e\u4e00\u5757\u5f3a\u52b2\u663e\u5361\uff1f</p>\n", 
      "date_published": "2020-04-09T06:39:15+00:00", 
      "title": "\u666e\u901a\u7684 Linux \u670d\u52a1\u5668\u600e\u4e48\u8dd1 tensorflow?", 
      "id": "https://www.v2ex.com/t/660798"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/0Nick", 
        "name": "0Nick", 
        "avatar": "https://cdn.v2ex.com/avatar/6be1/6790/435924_large.png?m=1572439601"
      }, 
      "url": "https://www.v2ex.com/t/602052", 
      "date_modified": "2019-09-19T02:26:54+00:00", 
      "content_html": "9 \u6708 22 \u65e5\uff0c\u4f5c\u4e3a\u56fd\u5e86\u524d\u7684\u897f\u5b89\u6700\u540e\u4e00\u573a TensorFlow \u6280\u672f\u4ea4\u6d41\uff0cTFUG \u5c06\u4e3a\u897f\u5b89\u5f00\u53d1\u8005\u5e26\u6765\u4e00\u573a\u5185\u5bb9\u4e30\u5bcc\u7684\u6df1\u5ea6\u5b66\u4e60\u6280\u672f\u4ea4\u6d41\u6d3b\u52a8\u3002\u5728\u672c\u6b21\u6d3b\u52a8\u4e0a\uff0c\u6765\u81ea\u4e0d\u540c\u4e13\u4e1a\u80cc\u666f\u7684\u4e09\u4f4d\u8bb2\u5e08\u5c06\u5206\u522b\u5c31\u673a\u5668\u5b66\u4e60\u9886\u57df\u8bed\u8a00\u65b0\u5ba0 Julia \uff08\u80e1\u952e\uff0c\u4e0a\u6d77\u572d\u6b65 CTO \uff09\u3001\u9886\u57df\u77e5\u8bc6\u56fe\u8c31\uff08\u9093\u53f6\u52cb\uff0c\u7f8e\u4e9a\u67cf\u79d1 AI \u7814\u53d1\u4e2d\u5fc3\u673a\u5668\u5b66\u4e60\u7b97\u6cd5\u5de5\u7a0b\u5e08\uff09\u548c TensorFlow \u5165\u95e8\uff08\u7fdf\u519b\u5fd7\uff0c\u6df1\u4fe1\u79d1\u521b\u9ad8\u7ea7\u5de5\u7a0b\u5e08\uff09\u7ed9\u5927\u5bb6\u5e26\u6765\u7cbe\u5f69\u7684\u5206\u4eab\u548c\u8ba8\u8bba\u3002<br />\u65f6\u95f4\uff1a9 \u6708 22 \u65e5 13:30-17:30<br />\u5730\u70b9\uff1a\u5929\u8c37\u516b\u8def\u8f6f\u4ef6\u65b0\u57ce 2 \u671f c1 \u963f\u91cc\u5df4\u5df4\u4e1d\u8def\u603b\u90e8 19 \u697c<br />\u516c\u4f17\u53f7\u641c\uff1aTFUG \u897f\u5b89", 
      "date_published": "2019-09-19T02:22:54+00:00", 
      "title": "\u897f\u5b89 TensorFlow \u6d3b\u52a8", 
      "id": "https://www.v2ex.com/t/602052"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/chinvo", 
        "name": "chinvo", 
        "avatar": "https://cdn.v2ex.com/avatar/7ff7/ca9c/42217_large.png?m=1682404407"
      }, 
      "url": "https://www.v2ex.com/t/600016", 
      "title": "\u81ea\u7136\u8bed\u8a00\u5904\u7406\uff0c\u5206\u8bcd\u95ee\u9898", 
      "id": "https://www.v2ex.com/t/600016", 
      "date_published": "2019-09-11T06:27:16+00:00", 
      "content_html": "<p>\u4eca\u5929\u5728 GDD \u542c\u4e86 tf.text \u4e3b\u9898\u6f14\u8bb2\uff0c\u6f14\u8bb2\u8005\u5728 demo \u4e2d\u4f7f\u7528 Unicode \u5206\u5b57\uff0c\u628a\u4e2d\u6587\u53e5\u5b50\u5206\u6210\u5355\u5b57\u3002<br/><br/>\u800c\u6211\u548c\u540c\u4e8b\u5728\u8fc7\u53bb\u591a\u7528\u8bcd\u5178\u6cd5\u5206\u8bcd\u3002<br/><br/>\u60f3\u8981\u8ba8\u8bba\u4e0b\u5206\u5b57\u548c\u8bcd\u5178\u6cd5\u5206\u8bcd\u4e24\u4e2a\u6548\u679c\u6709\u4ec0\u4e48\u5dee\u5f02\uff0c\u7ed3\u679c\u4e00\u76f4\u6ca1\u80fd\u5835\u5230\u6f14\u8bb2\u8005 \ud83d\ude02\u3002<br/><br/>\u4e0d\u77e5\u9053\u5404\u4f4d\u5728\u5e94\u7528\u8fc7\u7a0b\u4e2d\u6709\u6ca1\u6709\u5bf9\u8fd9\u65b9\u9762\u505a\u8fc7\u8bc4\u4f30\u3002</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/gancl", 
        "name": "gancl", 
        "avatar": "https://cdn.v2ex.com/avatar/757f/9aa4/53616_large.png?m=1401100020"
      }, 
      "url": "https://www.v2ex.com/t/592791", 
      "date_modified": "2019-08-17T16:12:41+00:00", 
      "content_html": "<p><a href=\"https://github.com/AKSHAYUBHAT/DeepVideoAnalytics\" rel=\"nofollow\">https://github.com/AKSHAYUBHAT/DeepVideoAnalytics</a>  \u6709\u627e\u4e86\u4e00\u4e2a,\u4f46\u662f\u90fd\u8dd1\u4e0d\u8d77\u6765,\u62a5 docker \u7684\u9519\u8bef.\n<a href=\"https://github.com/didi/delta\" rel=\"nofollow\">https://github.com/didi/delta</a>  \u8fd9\u4e2a\u6ef4\u6ef4\u5f00\u6e90\u7684,\u4f46\u662f\u53ea\u652f\u6301\u8bed\u97f3\u8bc6\u522b,\u8fd8\u4e0d\u652f\u6301\u56fe\u7247\u4eba\u8138\u8bc6\u522b.</p>\n<p>\u8fd8\u6709\u6ca1\u6709\u5176\u4ed6\u7684\u4f01\u4e1a\u53ef\u4ee5\u76f4\u63a5\u7528\u7684 ai \u5e73\u53f0?</p>\n<p>\u50cf\u817e\u8baf\u4e91\u3001\u963f\u91cc\u4e91\u7b49\u63d0\u4f9b api \u7684\u4e00\u662f\u4e0d\u5b89\u5168,\u4e8c\u662f\u8981\u6536\u8d39,\u81ea\u5efa\u4e00\u4e2a ai \u5e73\u53f0\u4f5c\u4e3a\u8fdb\u5165 ai \u9886\u57df\u7684\u7814\u53d1,\u8981\u600e\u4e48\u5165\u624b?</p>\n", 
      "date_published": "2019-08-17T16:12:12+00:00", 
      "title": "\u6709\u6ca1\u9002\u5408\u4f01\u4e1a\u5185\u90e8\u7528\u7684\u5f00\u6e90 AI \u5e73\u53f0?\u81ea\u5efa\u4e00\u4e2a ai \u5e73\u53f0\u4f5c\u4e3a\u8fdb\u5165 ai \u9886\u57df\u7684\u7814\u53d1,\u8981\u600e\u4e48\u5165\u624b?", 
      "id": "https://www.v2ex.com/t/592791"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/scalaer", 
        "name": "scalaer", 
        "avatar": "https://cdn.v2ex.com/avatar/2ad8/127f/268589_large.png?m=1757579908"
      }, 
      "url": "https://www.v2ex.com/t/587933", 
      "date_modified": "2019-07-31T08:57:18+00:00", 
      "content_html": "<pre><code class=\"language-shell\">| category      \t\t\t\t\t\n| -------------\t\t\t\t\t\t\t\n|'\u5bb6\u5c45', '\u5ba4\u5185\u8bbe\u8ba1', '\u88c5\u9970\u88c5\u6f62'\t\t\t  \n|'\u623f\u5730\u4ea7', '\u5efa\u7b51', '\u5efa\u6750', '\u5de5\u7a0b'   \t\t  \n|'\u623f\u5730\u4ea7'\n</code></pre>\n<p>\u4e0d\u652f\u6301\u8868\u683c\u3002\u3002\u3002</p>\n", 
      "date_published": "2019-07-31T08:56:47+00:00", 
      "title": "\u8bf7\u6559\u4e0b\uff0c \u7279\u5f81\u6709\u591a\u4e2a\u503c\uff0c \u600e\u4e48\u5904\u7406\u6bd4\u8f83\u597d\uff1f", 
      "id": "https://www.v2ex.com/t/587933"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/1722332572", 
        "name": "1722332572", 
        "avatar": "https://cdn.v2ex.com/avatar/31e3/2c4d/135473_large.png?m=1739072858"
      }, 
      "url": "https://www.v2ex.com/t/580826", 
      "date_modified": "2019-07-07T15:11:16+00:00", 
      "content_html": "<p>#Tensorflow \u5165\u95e8\u6559\u7a0b\n<a href=\"http://tensornews.cn/\" rel=\"nofollow\">http://tensornews.cn/</a></p>\n<ul>\n<li><a href=\"http://tensornews.cn/develop_deep_learning/\" rel=\"nofollow\">\u6df1\u5ea6\u5b66\u4e60\u53d1\u5c55\u53f2</a></li>\n<li><a href=\"http://tensornews.cn/feature_engineering/\" rel=\"nofollow\">\u7279\u5f81\u5de5\u7a0b</a></li>\n<li><a href=\"http://tensornews.cn/activation_function/\" rel=\"nofollow\">\u6df1\u5ea6\u5b66\u4e60\u4e4b\u6fc0\u6d3b\u51fd\u6570</a></li>\n<li><a href=\"http://tensornews.cn/loss_function/\" rel=\"nofollow\">\u635f\u5931\u51fd\u6570</a></li>\n<li><a href=\"http://tensornews.cn/backpropagation/\" rel=\"nofollow\">\u53cd\u5411\u4f20\u64ad\u7b97\u6cd5 [\u4e0a] </a></li>\n<li><a href=\"http://tensornews.cn/backpropagation_1/\" rel=\"nofollow\">\u53cd\u5411\u4f20\u64ad\u7b97\u6cd5 [\u4e0b] </a></li>\n<li><a href=\"http://tensornews.cn/introduction_and_install_tensorflow/\" rel=\"nofollow\">Tensorflow \u4ecb\u7ecd\u548c\u5b89\u88c5</a></li>\n<li><a href=\"http://tensornews.cn/tensorflow_basic/\" rel=\"nofollow\">Tensorflow \u57fa\u672c\u64cd\u4f5c</a></li>\n<li><a href=\"http://tensornews.cn/intro_cnn/\" rel=\"nofollow\">\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\u6982\u8ff0</a></li>\n<li><a href=\"http://tensornews.cn/cnn_implication/\" rel=\"nofollow\">\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\u57fa\u7840\u64cd\u4f5c</a></li>\n<li><a href=\"http://tensornews.cn/mnist_intro/\" rel=\"nofollow\">\u624b\u5199\u6570\u5b57\u8bc6\u522b\u6570\u636e\u96c6\u4ecb\u7ecd</a></li>\n<li><a href=\"http://tensornews.cn/mnist_implication/\" rel=\"nofollow\">\u5377\u79ef\u795e\u7ecf\u5b9e\u73b0\u624b\u5199\u6570\u5b57\u8bc6\u522b</a></li>\n<li><a href=\"http://tensornews.cn/cnn_models_intro/\" rel=\"nofollow\">\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\u7ecf\u5178\u6a21\u578b\u6982\u8ff0</a></li>\n<li><a href=\"http://tensornews.cn/lenet/\" rel=\"nofollow\">\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\u7ecf\u5178\u6a21\u578b LeNet</a></li>\n<li><a href=\"http://tensornews.cn/rnn_intro/\" rel=\"nofollow\">\u7efc\u8ff0 RNN \u5faa\u73af\u795e\u7ecf\u7f51\u7edc</a></li>\n<li><a href=\"http://tensornews.cn/rnn_implication/\" rel=\"nofollow\">\u4e03\u6b65\u5e26\u4f60\u5b9e\u73b0 RNN \u5faa\u73af\u795e\u7ecf\u7f51\u7edc\u5c0f\u793a\u4f8b</a></li>\n<li><a href=\"http://tensornews.cn/rnn_models_intro/\" rel=\"nofollow\">\u5e38\u7528 RNN \u7f51\u7edc\u7ed3\u6784\u53ca\u4f9d\u8d56\u4f18\u5316\u95ee\u9898</a></li>\n<li><a href=\"http://tensornews.cn/rnn_attention/\" rel=\"nofollow\">RNN \u7684\u5e94\u7528\u53ca\u6ce8\u610f\u529b\u6a21\u578b</a></li>\n<li><a href=\"http://tensornews.cn/Tensorboard_1/\" rel=\"nofollow\">Tensorboard  [\u4e0a] </a></li>\n<li><a href=\"http://tensornews.cn/Tensorboard_2/\" rel=\"nofollow\">Tensorboard  [\u4e0b] </a></li>\n<li><a href=\"http://tensornews.cn/multi_gpu/\" rel=\"nofollow\">tensorflow_multi_gpu</a></li>\n<li><a href=\"http://tensornews.cn/Awesome_Tensorflow/\" rel=\"nofollow\">Awesome_Tensorflow</a></li>\n</ul>\n", 
      "date_published": "2019-07-07T15:10:52+00:00", 
      "title": "Tensorflow \u5165\u95e8\u6559\u7a0b", 
      "id": "https://www.v2ex.com/t/580826"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/different", 
        "name": "different", 
        "avatar": "https://cdn.v2ex.com/avatar/94f6/7670/374456_large.png?m=1661142546"
      }, 
      "url": "https://www.v2ex.com/t/578357", 
      "title": "\u5173\u4e8e spyder3 \u7684 Python \u4ee3\u7801\u8fd0\u884c\u7ed3\u679c\u4e0d\u4e00\u81f4", 
      "id": "https://www.v2ex.com/t/578357", 
      "date_published": "2019-06-28T08:24:24+00:00", 
      "content_html": "<p>\u540c\u6837\u7684\u4ee3\u7801,\u5728 spyder \u91cc\u9762\u8fd0\u884c\u548c\u76f4\u63a5\u5728 Ubuntu \u63a7\u5236\u53f0\u4e0a\u8fd0\u884c\u7684\u7ed3\u679c\u4e0d\u4e00\u81f4,\u5982\u4e0b\u622a\u56fe\nspyder \u7684\u63a7\u5236\u53f0:<a href=\"https://i.loli.net/2019/06/28/5d15cc7f0daf243402.png\" rel=\"nofollow\">https://i.loli.net/2019/06/28/5d15cc7f0daf243402.png</a>\nubuntu \u63a7\u5236\u53f0:<a href=\"https://i.loli.net/2019/06/28/5d15cc7fcf86410820.png\" rel=\"nofollow\">https://i.loli.net/2019/06/28/5d15cc7fcf86410820.png</a></p>\n<p>\u987a\u4fbf\u95ee\u4e00\u53e5,tensorflow,\u4e00\u76f4\u6ca1\u6709\u4f7f\u7528 GPU \u6765\u8dd1,\u8fd9\u662f\u4e3a\u5565?\ncuda \u7248\u672c:10.01\npython:3.6.7\ntensorflow                    1.13.1<br>\ntensorflow-estimator          1.13.0<br>\ntensorflow-gpu                1.13.1\ncudnn 7.5\n\u4e5f\u8bd5\u8fc7\u5c06\u8fd9\u4e9b\u4e1c\u897f\u91cd\u88c5\u4e86\u4e00\u6b21,\u8fd8\u662f\u4e0d\u884c\n\u8fd0\u884c tensorflow \u7684\u65f6\u5019,\u4e00\u76f4\u6ca1\u6709\u62a5\u9519,\u4f46\u662f\u628a\u6211\u7684\u663e\u5361\u5185\u5b58\u5360\u7528\u4e86,\u5374\u5728\u7528 cpu \u6765\u8dd1,GPU(GTX960)\u5229\u7528\u7387\u4e3a&lt;10%\n\u90e8\u5206\u4ee3\u7801:\ngpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.33)\nwith tf.Session(gpu_options=gpu_options,config=tf.ConfigProto(log_device_placement=True)) as sess:\nsess.run(mymode.global_init)                         #\u521d\u59cb\u5168\u5c40\u8282\u70b9\n\u90c1\u95f7...</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/LuJyKa", 
        "name": "LuJyKa", 
        "avatar": "https://cdn.v2ex.com/avatar/8273/0d4b/380379_large.png?m=1759229110"
      }, 
      "url": "https://www.v2ex.com/t/574840", 
      "title": "\u73b0\u5728\u662f\u4e0d\u662f\u8f6c MXNET \u7684\u6700\u597d\u65f6\u673a", 
      "id": "https://www.v2ex.com/t/574840", 
      "date_published": "2019-06-17T11:41:56+00:00", 
      "content_html": "<p>\u7528\u4e86\u65b0\u7684 Tesorflow\uff0c\u611f\u89c9\u5b8c\u5168\u4e3a\u4e86\u7167\u987e\u5c0f\u767d\uff0c\u628a\u539f\u6765\u7684\u8001\u7528\u6237\u7ed9\u5c60\u6740\u7684\u5e72\u5e72\u51c0\u51c0\u3002\n\u52a8\u6001\u56fe\u8ba1\u7b97\uff0c\u6211\u6839\u672c\u4e0d\u9700\u8981\uff0c\u6027\u80fd\u3001\u79fb\u690d\u6027\u662f\u6211\u8003\u8651\u7684\u76ee\u6807\u3002\n\u73b0\u5728 TF \u4e0d\u4ec5\u5305\u8d8a\u6765\u8d8a\u5927\uff0c\u8d8b\u4e8e\u81c3\u80bf\uff0c\u8fde\u8bed\u6cd5\u90fd\u770b\u4e0d\u61c2\uff0c\u53eb TfKeras \u5f97\u4e86\u3002\n\u6709\u4ecb\u7ecd\u8bf4 MXNET \u5f88\u597d\uff0c\u770b\u770b\u5927\u5bb6\u600e\u4e48\u60f3\u3002</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/lunafreya", 
        "name": "lunafreya", 
        "avatar": "https://cdn.v2ex.com/gravatar/0abb60fa4c6e891ef6ec8b25df81fc19?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/561011", 
      "title": "\u5bf9\u4e8e TensorFlow 2.0 \u7684 Keras API\uff0c \u5927\u5bb6\u600e\u4e48\u770b", 
      "id": "https://www.v2ex.com/t/561011", 
      "date_published": "2019-05-05T02:29:45+00:00", 
      "content_html": "<p>\u5176\u5b9e\u73b0\u5728 TensorFlow 1.13 \u7248\u5df2\u7ecf\u6709\u5927\u91cf tf.keras \u7684 API \u4e86\uff0c \u6211\u5c1d\u8bd5\u5c06 MobileNet V2 \u7528 keras \u90a3\u5957 subclassing \u65b9\u6cd5\u5199\u4e86\u4e00\u904d\u3002 \u53d1\u73b0\u65e0\u8bba\u600e\u6837\u64cd\u4f5c\u90fd\u65e0\u6cd5\u628a tf.keras.layers \u91cc Batch Normalization \u7684 running mean \u4ee5\u53ca running variance \u52a0\u8fdb UpdateOps \u8fd9\u4e2a Collection, \u540c\u65f6\uff0c\u7c7b\u4f3c\u8fd9\u79cd\u60c5\u51b5\u542b\u6709 regularization \u7684 Conv2D \u4e2d\u4e5f\u4e0d\u4f1a\u628a regularization \u52a0\u8fdb regularization loss collection \u91cc\u3002 \u67e5\u4e86\u4e0b GitHub \u4e0a\u76f8\u5173\u7684 issue, \u6709\u4eba\u8bf4\u7528 keras \u90a3\u5957 model.compile \u548c model.fit \u6765\u8fdb\u884c training \u53ef\u4ee5 update \u8fd9\u4e9b\u3002 \u8fd9\u662f\u4e0d\u662f\u610f\u5473\u7740\u4ee5\u540e\u4e60\u60ef\u9759\u6001\u56fe\u7684\u73a9\u5bb6\u57fa\u672c\u5c31\u6ca1\u6709\u6846\u67b6\u53ef\u7528\u4e86\u3002</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/Zcalory", 
        "name": "Zcalory", 
        "avatar": "https://cdn.v2ex.com/avatar/65bd/53d2/392963_large.png?m=1554796863"
      }, 
      "url": "https://www.v2ex.com/t/554568", 
      "title": "[Tensorboard \u95ee\u9898] \u5728\u5df2\u6dfb\u52a0 tf.summary.scalar()\u7684\u60c5\u51b5\u4e0b\uff0c tf.summary.merge_all() \u8fd4\u56de\u4e3a None. \r\n\u8bf7\u95ee\u8fd9\u79cd\u60c5\u51b5\u4e3a\u4f55\u53d1\u751f\uff0c\u5982\u4f55\u89e3\u51b3\uff1f", 
      "id": "https://www.v2ex.com/t/554568", 
      "date_published": "2019-04-12T10:02:58+00:00", 
      "content_html": "<p>\u88ab Tensorboard \u6298\u78e8 ing, \u78b0\u5230\u8fd9\u4e2a\u95ee\u9898\uff0c\u4e0d\u5f97\u5176\u89e3\u3002\n\u9996\u5148\u6211\u7528 placeholder \u5b9a\u4e49\u4e86\u4e00\u4e2a\u8ba1\u7b97\u56fe\uff0c\u5e76\u5c06\u5176\u4e2d\u7684\u67d0\u4e9b\u53d8\u91cf\u6dfb\u52a0\u5230\u4e86 tf.summary.scalar()\u4e2d\u3002\n\u968f\u540e\u6211\u4ee5\u5982\u4e0b\u65b9\u5f0f\u6dfb\u52a0\u4e86 tf.summary.merge_all()</p>\n<pre><code>    # Graph has been defined above\n    # And variables have beed added via tf.summary.scalar()\n    init = tf.global_variables_initializer()\n    if save_dir is not None:\n        merged_summary_op = tf.summary.merge_all()\n\n    # Training process...\n    with tf.Session() as sess:\n        sess.run(init)\n        print(tf.get_collection(tf.GraphKeys.SUMMARIES), merged_summary_op)\n        # returns []  None\n</code></pre>\n<p>\u62a5\u9519\u663e\u793a\u4e3a\uff1a</p>\n<pre><code>        cost, rcost, merged = sess.run([self.cost, self.rcost, merged_summary_op], feed_dict={self.mode:'train', self.x:data, self.labels:label})\n        TypeError: Fetch argument None has invalid type &lt;class 'NoneType'&gt; \uff08 merged_summary_op -&gt; None \uff09\n</code></pre>\n<p>\u8bf7\u95ee\u8fd9\u662f\u4ec0\u4e48\u539f\u56e0\u5bfc\u81f4\u7684\uff1f\n\u6211\u67e5\u5230\u4e86\u4e00\u7bc7\u535a\u5ba2\uff0c\u91cc\u9762\u63cf\u8ff0\u4e86\u4e00\u79cd\u60c5\u5f62\uff0c\u6211\u4e5f\u4e0d\u592a\u7406\u89e3\u3002\u8fd9\u4f1a\u662f\u8ddf\u6211\u7c7b\u4f3c\u7684\u95ee\u9898\u5417\uff1f<br>\n<a href=\"https://blog.csdn.net/hu_guan_jie/article/details/78477117\" rel=\"nofollow\">\u535a\u5ba2\u94fe\u63a5</a></p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/KarlRixon", 
        "name": "KarlRixon", 
        "avatar": "https://cdn.v2ex.com/avatar/ed12/aad2/365336_large.png?m=1550115885"
      }, 
      "url": "https://www.v2ex.com/t/538115", 
      "date_modified": "2019-02-24T02:46:47+00:00", 
      "content_html": "<p>\u6570\u636e\u6765\u6e90\u662f\u4eac\u4e1c\u65b0\u6b3e\u624b\u673a\u7684\u8bc4\u8bba\u548c\u6253\u5206\uff0c\u76ee\u524d\u6536\u96c6\u5230 2500 \u6761\u6570\u636e\uff0c\u4f46\u6253\u5206\u5c0f\u4e8e 5 \u5206\u7684\u53ea\u6709\u4e0d\u5230 40 \u6761</p>\n<p>\u8bad\u7ec3\u6a21\u578b\u5c42\u6b21\u662f\uff1a\u5d4c\u5165\u5c42-\u300b LSTM-\u300b Dense-\u300b Dense-\u300b\u8f93\u51fa\u5c42\n\u5d4c\u5165\u5c42\u7684\u521d\u59cb\u6570\u636e\u4e3a word2vec \u8bad\u7ec3\u7684\u8bcd\u5411\u91cf\n\u8f93\u5165\u7684\u8bad\u7ec3\u6570\u636e\u4e3a\u8bcd\u7d22\u5f15\uff0c\u6807\u8bb0\u4e3a\u6253\u5206</p>\n<p>\u90e8\u5206\u4ee3\u7801\u5982\u4e0b\uff1a</p>\n<pre><code class=\"language-python\">def main():\n    x_train = pad_seq()\n    y_train = star()\n    x_train, y_train, x_test, y_test = set_data(x_train, y_train)\n    model = Sequential()\n    model.add(Embedding(input_dim=input_dim+1, output_dim=output_dim, input_length=k, embeddings_initializer=my_init))\n    model.add(LSTM(128, dropout=0.2, recurrent_dropout=0.2, activation='sigmoid'))\n    model.add(Dense(256, activation='relu'))\n    model.add(Dense(128, activation='relu'))\n    model.add(Dense(1,activation='sigmoid'))\n    model.compile(loss=\"binary_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])\n    model.fit(x_train, y_train, batch_size=batch_size, epochs=15)\n    score, acc = model.evaluate(x_test, y_test, batch_size=batch_size)\n    print('Test score:', score)\n    print('Test accuracy:', acc)\n</code></pre>\n<p>\u7136\u800c\u8bad\u7ec3\u7ed3\u679c\u662f\u8fd9\u6837\u7684\u3002\u3002\u3002</p>\n<pre><code class=\"language-cmd\">Epoch 15/15\n\n  32/2015 [..............................] - ETA: 0s - loss: -63.2713 - acc: 0.0000e+00\n 160/2015 [=&gt;............................] - ETA: 0s - loss: -63.4706 - acc: 0.0000e+00\n 288/2015 [===&gt;..........................] - ETA: 0s - loss: -63.5481 - acc: 0.0000e+00\n 384/2015 [====&gt;.........................] - ETA: 0s - loss: -63.2713 - acc: 0.0026    \n 480/2015 [======&gt;.......................] - ETA: 0s - loss: -63.3046 - acc: 0.0021\n 608/2015 [========&gt;.....................] - ETA: 0s - loss: -63.4024 - acc: 0.0016\n 736/2015 [=========&gt;....................] - ETA: 0s - loss: -63.3796 - acc: 0.0027\n 864/2015 [===========&gt;..................] - ETA: 0s - loss: -63.3821 - acc: 0.0023\n 992/2015 [=============&gt;................] - ETA: 0s - loss: -63.3838 - acc: 0.0020\n1120/2015 [===============&gt;..............] - ETA: 0s - loss: -63.3852 - acc: 0.0018\n1248/2015 [=================&gt;............] - ETA: 0s - loss: -63.3991 - acc: 0.0016\n1376/2015 [===================&gt;..........] - ETA: 0s - loss: -63.4104 - acc: 0.0015\n1504/2015 [=====================&gt;........] - ETA: 0s - loss: -63.3879 - acc: 0.0020\n1632/2015 [=======================&gt;......] - ETA: 0s - loss: -63.4081 - acc: 0.0018\n1760/2015 [=========================&gt;....] - ETA: 0s - loss: -63.4344 - acc: 0.0017\n1888/2015 [===========================&gt;..] - ETA: 0s - loss: -63.4233 - acc: 0.0021\n2015/2015 [==============================] - 1s 470us/step - loss: -63.4214 - acc: 0.0020\n\n 32/504 [&gt;.............................] - ETA: 2s\n504/504 [==============================] - 0s 412us/step\nTest score: -63.769539061046785\nTest accuracy: 0.0\n\nProcess finished with exit code 0\n</code></pre>\n<p>model \u5982\u4e0b\uff1a</p>\n<pre><code class=\"language-cmd\">_________________________________________________________________\nLayer (type)                 Output Shape              Param #   \n=================================================================\nembedding_1 (Embedding)      (None, 20, 50)            49350     \n_________________________________________________________________\nlstm_1 (LSTM)                (None, 128)               91648     \n_________________________________________________________________\ndense_1 (Dense)              (None, 256)               33024     \n_________________________________________________________________\ndense_2 (Dense)              (None, 128)               32896     \n_________________________________________________________________\ndense_3 (Dense)              (None, 1)                 129       \n=================================================================\nTotal params: 207,047\nTrainable params: 207,047\nNon-trainable params: 0\n_________________________________________________________________\n</code></pre>\n", 
      "date_published": "2019-02-24T02:42:10+00:00", 
      "title": "word2vec+LSTM \u60c5\u611f\u8bc6\u522b\uff0c\u5e2e\u5fd9\u770b\u770b\u54ea\u51fa\u4e86\u95ee\u9898", 
      "id": "https://www.v2ex.com/t/538115"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/KarlRixon", 
        "name": "KarlRixon", 
        "avatar": "https://cdn.v2ex.com/avatar/ed12/aad2/365336_large.png?m=1550115885"
      }, 
      "url": "https://www.v2ex.com/t/534851", 
      "title": "\u5b89\u88c5 tensorflow \u7684 gpu \u7248\u540e import \u62a5\u9519\uff0c\u6709\u5927\u4f6c\u78b0\u5230\u8fc7\u8fd9\u79cd\u60c5\u51b5\u5417\uff1f", 
      "id": "https://www.v2ex.com/t/534851", 
      "date_published": "2019-02-14T03:45:48+00:00", 
      "content_html": "<p>\u9996\u5148 cuda \u88c5\u7684\u662f 9.1.85_win10_64 \u4f4d\uff0c\u662f\u5728\u767e\u5ea6\u7f51\u76d8\u4e0a\u4e0b\u8f7d\u7684 local \u53ef\u6267\u884c\u6587\u4ef6\uff08\u56e0\u4e3a\u5b98\u7f51\u7684\u4e0b\u8f7d\u592a\u6162\u800c\u4e14\u8054\u7f51\u7248\u5b89\u88c5\u5305\u4e5f\u5f88\u6162\uff09\uff0c\u7528 vs2015 \u6d4b\u8bd5\u81ea\u5e26 Samples \u6210\u529f\u3002\n<br>\n\u7136\u540e cudnn \u662f\u4e0b\u8f7d\u5b98\u7f51\u7684 9.0 \u7248\u672c\uff0c\u4f46\u662f\u6309\u7167\u7f51\u4e0a\u505a\u6cd5\u6d4b\u8bd5\u62a5\u9519\uff1a</p>\n<pre><code>#include &lt;iostream&gt;\n#include &lt;cuda_runtime.h&gt;\n#include &lt;cudnn.h&gt;\nusing namespace std;\n\nvoid main() {\n\tcudnnHandle_t handle;\n\tcudnnStatus_t t = cudnnCreate(&amp;handle);\n\tcout &lt;&lt; cudnnGetErrorString(t);\n\tgetchar();\n}\n</code></pre>\n<p>\u9519\u8bef\tMSB3721\t\u547d\u4ee4\u201c\"D:\\CUDA9.2\\Development\\bin\\nvcc.exe\" -gencode=arch=compute_30,code=\"sm_30,compute_30\" --use-local-env --cl-version 2015 -ccbin \"C:\\Program Files (x86)\\Microsoft Visual Studio 14.0\\VC\\bin\" -x cu  -ID:\\CUDA9.2\\Development\\include -ID:\\CUDA9.2\\Development\\include  -G   --keep-dir Debug -maxrregcount=0  --machine 32 --compile -cudart static  -g   -D_MBCS -Xcompiler \"/EHsc /W3 /nologo /Od /FS /Zi /RTC1 /MDd \" -o Debug\\test.cu.obj \"C:\\Users\\24346\\Documents\\C++\\testCUDA\\test_cudnn\\test_cudnn\\<a href=\"http://test.cu\" rel=\"nofollow\">test.cu</a>\"\u201d\u5df2\u9000\u51fa\uff0c\u8fd4\u56de\u4ee3\u7801\u4e3a 1\u3002\n<br><br>\nMicrosoft Visual C++ 2017 Redistributable(X64)\u5df2\u5b89\u88c5\n<br>\npython \u662f 3.6.8 \u7248\u672c\u7684 64 \u4f4d\uff0cpip install tensorflow-gpu \u663e\u793a\u5b89\u88c5\u6210\u529f\uff0c\u4f46 import tensorflow \u62a5\u9519\uff1a</p>\n<pre><code>ImportError: Traceback (most recent call last):\n  File \"D:\\anaconda\\envs\\tensorflow-gpu\\lib\\site-packages\\tensorflow\\python\\pywrap_tensorflow_internal.py\", line 18, in swig_import_helper\n    return importlib.import_module(mname)\n  File \"D:\\anaconda\\envs\\tensorflow-gpu\\lib\\importlib\\__init__.py\", line 126, in import_module\n    return _bootstrap._gcd_import(name[level:], package, level)\n  File \"&lt;frozen importlib._bootstrap&gt;\", line 978, in _gcd_import\n  File \"&lt;frozen importlib._bootstrap&gt;\", line 961, in _find_and_load\n  File \"&lt;frozen importlib._bootstrap&gt;\", line 950, in _find_and_load_unlocked\n  File \"&lt;frozen importlib._bootstrap&gt;\", line 648, in _load_unlocked\n  File \"&lt;frozen importlib._bootstrap&gt;\", line 560, in module_from_spec\n  File \"&lt;frozen importlib._bootstrap_external&gt;\", line 922, in create_module\n  File \"&lt;frozen importlib._bootstrap&gt;\", line 205, in _call_with_frames_removed\nImportError: DLL load failed: \u627e\u4e0d\u5230\u6307\u5b9a\u7684\u6a21\u5757\u3002\n\nDuring handling of the above exception, another exception occurred:\n\nTraceback (most recent call last):\n  File \"D:\\anaconda\\envs\\tensorflow-gpu\\lib\\site-packages\\tensorflow\\python\\pywrap_tensorflow.py\", line 41, in &lt;module&gt;\n    from tensorflow.python.pywrap_tensorflow_internal import *\n  File \"D:\\anaconda\\envs\\tensorflow-gpu\\lib\\site-packages\\tensorflow\\python\\pywrap_tensorflow_internal.py\", line 21, in &lt;module&gt;\n    _pywrap_tensorflow_internal = swig_import_helper()\n  File \"D:\\anaconda\\envs\\tensorflow-gpu\\lib\\site-packages\\tensorflow\\python\\pywrap_tensorflow_internal.py\", line 20, in swig_import_helper\n    return importlib.import_module('_pywrap_tensorflow_internal')\n  File \"D:\\anaconda\\envs\\tensorflow-gpu\\lib\\importlib\\__init__.py\", line 126, in import_module\n    return _bootstrap._gcd_import(name[level:], package, level)\nModuleNotFoundError: No module named '_pywrap_tensorflow_internal'\n\n\nFailed to load the native TensorFlow runtime.\n\nSee https://www.tensorflow.org/install/install_sources#common_installation_problems\n\nfor some common reasons and solutions.  Include the entire stack trace\nabove this error message when asking for help.\n</code></pre>\n<p>\u4f7f\u7528 anaconda \u5b89\u88c5 tensorflow-gpu \u4e5f\u51fa\u73b0\u8fd9\u4e2a ImportError\n<br>\n\u73b0\u5728\u6211\u5b89\u88c5\u4e86 tensorflow \u7684 cpu \u7248\u5c31\u6ca1\u6709\u8fd9\u4e2a\u95ee\u9898\u4e86\u3002\u3002\u3002\n<br>\n\u60f3\u95ee\u4e00\u4e0b 10000 \u6761\u4ee5\u5185\u8bc4\u8bba\u7684\u8bcd\u5411\u91cf\u8bad\u7ec3\u9700\u8981\u663e\u5361\u52a0\u901f\u5417</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/Ewig", 
        "name": "Ewig", 
        "avatar": "https://cdn.v2ex.com/gravatar/4f00a030a1532d147ba8ec51ef52ad26?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/526713", 
      "date_modified": "2019-01-13T16:06:41+00:00", 
      "content_html": "\u6211\u7684 python \u7248\u672c\u662f Python 3.4.8<br /><br />shenjianlin@newdev:~(spider)$ pip install --ignore-installed --upgrade '<a target=\"_blank\" href=\"https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl'--user\" rel=\"nofollow\">https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl'--user</a><br />Collecting <a target=\"_blank\" href=\"https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user\" rel=\"nofollow\">https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user</a><br />  HTTP error 404 while getting <a target=\"_blank\" href=\"https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user\" rel=\"nofollow\">https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user</a><br />  Could not install requirement <a target=\"_blank\" href=\"https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user\" rel=\"nofollow\">https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user</a> because of error 404 Client Error: Not Found for url: <a target=\"_blank\" href=\"https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user\" rel=\"nofollow\">https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user</a><br />Could not install requirement <a target=\"_blank\" href=\"https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user\" rel=\"nofollow\">https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user</a> because of HTTP error 404 Client Error: Not Found for url: <a target=\"_blank\" href=\"https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user\" rel=\"nofollow\">https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user</a> for URL <a target=\"_blank\" href=\"https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user\" rel=\"nofollow\">https://github.com/sigilioso/tensorflow-build/raw/master/tensorflow-1.4.0-cp36-cp36m-linux_x86_64.whl--user</a><br /><br />\u4e0b\u8f7d\u7248\u672c\u7684\u62a5\u9519\uff0c\u76f4\u63a5 pip3 install tensorflow \u7684\u8bdd\u4e0d\u80fd\u7528<br /><br />Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA<br /><br />\u6211\u662f\u60f3\u7528\u8fd9\u4e2a\u6d6e\u70b9\u8fd0\u7b97\u7684\uff0c\u73b0\u5728\u5982\u4f55\u4e0b\u8f7d\u5bf9\u5e94\u7248\u672c\u7684 tensorflow", 
      "date_published": "2019-01-13T16:06:34+00:00", 
      "title": "tensorflow \u5b89\u88c5\u62a5\u9519", 
      "id": "https://www.v2ex.com/t/526713"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/suley", 
        "name": "suley", 
        "avatar": "https://cdn.v2ex.com/avatar/0628/fc47/31544_large.png?m=1704718794"
      }, 
      "url": "https://www.v2ex.com/t/526433", 
      "date_modified": "2019-01-12T14:38:09+00:00", 
      "content_html": "<p>\u6211\u662f\u53c2\u8003\u8fd9\u4f4d\u5144\u5f1f\u7684\u6587\u7ae0\u7ec3\u4e60\u7684\uff0c\u53ea\u4e0d\u8fc7\u6211\u7684\u73af\u5883\u6362\u6210\u4e86 Ubuntu\uff0c</p>\n<p><a href=\"https://www.jianshu.com/p/db8824205fc3\" rel=\"nofollow\">https://www.jianshu.com/p/db8824205fc3</a></p>\n<p>\u6839\u636e\u8fd9\u4f4d\u535a\u4e3b\u7684\u6d4b\u8bd5\uff0c\u7b2c\u4e00\u5343\u6b65\u5c31\u5dee\u4e0d\u591a\u80fd\u8fbe\u5230 90%\u51c6\u786e\u7387\u3002</p>\n<pre><code>step:100 loss:1.5357 accuracy:0.4900\nstep:200 loss:1.0189 accuracy:0.7000\nstep:300 loss:0.7720 accuracy:0.7850\nstep:400 loss:0.6589 accuracy:0.8000\nstep:500 loss:0.4987 accuracy:0.8300\nstep:600 loss:0.5863 accuracy:0.8100\nstep:700 loss:0.5242 accuracy:0.8350\nstep:800 loss:0.3541 accuracy:0.9000\nstep:900 loss:0.5004 accuracy:0.8700\nstep:1000 loss:0.3152 accuracy:0.9050\n</code></pre>\n<p>\u53ef\u662f\u6211\u8bad\u7ec3\u4e86\u5dee\u4e0d\u591a\u4e00\u767e\u4e07\u6b65\uff0c\u59cb\u7ec8\u5728 0.3~0.4 \u5f98\u5f8a\uff0c\u4e0d\u6536\u655b\uff0c\u91cc\u9762\u6709\u4e9b\u5927\u5c0f\u5199\u7684\u9519\u8bef\u6211\u90fd\u4fee\u6b63\u4e86\uff0c\u8fd8\u662f\u4e0d\u5f97\u8981\u9886\u3002</p>\n<p>\u6211\u7684\u8bad\u7ec3\u7ed3\u679c\uff1a</p>\n<pre><code>step:100 loss:2.2414 accuracy:0.2250 used time: 3 s\nstep:200 loss:2.0146 accuracy:0.3450 used time: 7 s\nstep:300 loss:1.9829 accuracy:0.2900 used time: 11 s\nstep:400 loss:1.8127 accuracy:0.3350 used time: 15 s\nstep:500 loss:1.9361 accuracy:0.3150 used time: 19 s\nstep:600 loss:1.8108 accuracy:0.3300 used time: 23 s\nstep:700 loss:1.7482 accuracy:0.3950 used time: 27 s\nstep:800 loss:1.7227 accuracy:0.3200 used time: 31 s\nstep:900 loss:1.7529 accuracy:0.3500 used time: 35 s\nstep:1000 loss:1.7124 accuracy:0.3300 used time: 38 s\nstep:1100 loss:1.7832 accuracy:0.3350 used time: 42 s\nstep:1200 loss:1.7278 accuracy:0.3500 used time: 46 s\nstep:1300 loss:1.6402 accuracy:0.3000 used time: 50 s\nstep:1400 loss:1.6699 accuracy:0.3200 used time: 54 s\nstep:1500 loss:1.6819 accuracy:0.3600 used time: 59 s\nstep:1600 loss:1.7417 accuracy:0.3400 used time: 63 s\nstep:1700 loss:1.7227 accuracy:0.3350 used time: 67 s\nstep:1800 loss:1.6762 accuracy:0.3850 used time: 71 s\nstep:1900 loss:1.6828 accuracy:0.3150 used time: 75 s\nstep:2000 loss:1.6694 accuracy:0.2900 used time: 79 s\nstep:2100 loss:1.6974 accuracy:0.2950 used time: 83 s\nstep:2200 loss:1.6517 accuracy:0.3450 used time: 87 s\nstep:2300 loss:1.6009 accuracy:0.3600 used time: 91 s\nstep:2400 loss:1.7358 accuracy:0.3300 used time: 95 s\nstep:2500 loss:1.7149 accuracy:0.3500 used time: 99 s\nstep:2600 loss:1.6166 accuracy:0.3850 used time: 103 s\nstep:2700 loss:1.6242 accuracy:0.3500 used time: 107 s\nstep:2800 loss:1.6648 accuracy:0.3550 used time: 111 s\nstep:2900 loss:1.6295 accuracy:0.3050 used time: 115 s\nstep:3000 loss:1.6616 accuracy:0.3400 used time: 119 s\n</code></pre>\n<p>\u8054\u7cfb\u4e86\u535a\u4e3b\uff0c\u535a\u4e3b\u8868\u793a\u68c0\u67e5\u4e86\u5f88\u4e45\u4e5f\u4e0d\u77e5\u9053\u4e3a\u4ec0\u4e48\u4f1a\u8fd9\u6837\u3002\u800c\u4e14\u540c\u4e00\u7bc7\u6587\u7ae0\u4e0b\u4e5f\u6709\u5176\u5b83\u7f51\u53cb\u8bf4\u9047\u5230\u4e86\u540c\u6837\u7684\u95ee\u9898\u3002</p>\n<p>\u8bf7\u6559\u4e0b\u5404\u4f4d\u5927\u795e~\u5148\u8c22\u8fc7\u4e86\u3002</p>\n", 
      "date_published": "2019-01-12T13:44:49+00:00", 
      "title": "\u8bf7\u6559: \u540c\u4e00\u4efd Tensorflow \u4ee3\u7801\uff0c\u8bad\u7ec3\u65b0\u95fb\u81ea\u52a8\u5206\u7c7b\uff0c Windows \u548c Linux \u4e0b\u7684\u8bad\u7ec3\u7ed3\u679c\u4e0d\u4e00\u81f4\uff1f", 
      "id": "https://www.v2ex.com/t/526433"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/nezzzz", 
        "name": "nezzzz", 
        "avatar": "https://cdn.v2ex.com/avatar/b323/c538/337723_large.png?m=1654583990"
      }, 
      "url": "https://www.v2ex.com/t/526404", 
      "date_modified": "2020-07-14T17:26:55+00:00", 
      "content_html": "<p>\u6700\u8fd1\u5728\u5b66\u4e60 python\uff0c\u60f3\u5165\u95e8\u4e00\u4e0b\u6df1\u5ea6\u5b66\u4e60\uff0c\u7f51\u4e0a\u641c\u4e86\u4e00\u4e9b\u5927\u90fd\u662f\u7528\u7684 python2.7\uff0c\u56e0\u4e3a\u521a\u5f00\u59cb\u6240\u4ee5\u4e0d\u592a\u660e\u767d\uff0cTensorFlow \u662f\u652f\u6301 python3 \u7684\uff0c\u8bf7\u6559\u5404\u4f4d\u4e00\u4e0b\uff0c\u6709\u597d\u7684\u4e66\u7c4d\u6216\u8005\u89c6\u9891\u63a8\u8350\u5417\uff1f</p>\n", 
      "date_published": "2019-01-12T10:55:52+00:00", 
      "title": "\u8bf7\u5404\u4f4d\u63a8\u8350 TensorFlow \u5165\u95e8\u5b66\u4e60\u65b9\u6cd5\u6216\u8def\u5f84", 
      "id": "https://www.v2ex.com/t/526404"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/redleaf", 
        "name": "redleaf", 
        "avatar": "https://cdn.v2ex.com/gravatar/644d35be6dcffbf917f1593d081419cc?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/519628", 
      "title": "\u5f00\u7bb1\u5373\u7528\u7684 Manjaro \u5927\u6cd5\u597d", 
      "id": "https://www.v2ex.com/t/519628", 
      "date_published": "2018-12-21T02:06:05+00:00", 
      "content_html": "<p>\u9ad8\u6df1\u5ea6\u5b66\u4e60\uff0c\u5b9e\u5728\u5fcd\u53d7\u4e0d\u4e86 windows \u4e0b\u4e00\u4e9b\u83ab\u540d\u5176\u5999\u7684\u9519\u8bef\uff0c\u6628\u5929\u6295\u9760\u4e86 Manjaro</p>\n<p>\u4ece\u5b89\u88c5\u5230\u4f7f\u7528\u7279\u522b\u7b80\u5355\uff0c\u76f4\u63a5\u5728 Manjaro \u81ea\u5e26\u7684\u8f6f\u4ef6\u5b89\u88c5\u5de5\u5177\u4e2d\u5b89\u88c5\u4e86 tensorflow-cuda\uff0c\u5b8c\u7f8e\u4f7f\u7528</p>\n<p>\u8bf4\u8bf4\u76f8\u6bd4\u4e8e pip \u5b89\u88c5\u7684\u51e0\u70b9\u597d\u5904</p>\n<ul>\n<li>pip \u4e2d\u7684 TensorFlow \u4e0d\u652f\u6301 python 3.7\uff0c\u800c Manjaro \u8f6f\u4ef6\u6e90\u4e2d\u7684\u662f\u9488\u5bf9 3.7 \u7f16\u8bd1\u7684</li>\n<li>pip \u4e2d\u7684 TensorFlow \u9700\u8981\u642d\u914d\u6307\u5b9a\u7248\u672c\u7684 CUDA \u548c cuDNN\uff0cManjaro \u4e2d\u8ddf CUDA \u548c cuDNN \u7248\u672c\u81ea\u52a8\u5339\u914d\uff08\u8c8c\u4f3c\u5df2\u7ecf\u662f CUDA10 \u4e86</li>\n<li>\u771f\u00b7\u96f6\u914d\u7f6e</li>\n</ul>\n<p>\u6628\u665a\u552f\u4e00\u6ca1\u89e3\u51b3\u7684\u5c31\u662f\u8f93\u5165\u6cd5\u95ee\u9898\uff0c\u7cfb\u7edf\u6258\u76d8\u4e2d\u7684\u8f93\u5165\u6cd5\u56fe\u6807\u51fa\u6765\u4e86\uff0c\u4f46\u662f\u8fd8\u4e0d\u80fd\u5207\u6362\u5230\u641c\u72d7\u8f93\u5165\u6cd5 00</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/gancl", 
        "name": "gancl", 
        "avatar": "https://cdn.v2ex.com/avatar/757f/9aa4/53616_large.png?m=1401100020"
      }, 
      "url": "https://www.v2ex.com/t/507194", 
      "title": "ai \u7684\u5728\u533b\u7597\u884c\u4e1a\u7684\u5e94\u7528\u6709\u54ea\u4e9b?\u6709\u54ea\u4e9b\u5f00\u6e90\u7684\u6216\u4e91\u670d\u52a1\u53ef\u4ee5\u76f4\u63a5\u7528\u5417?", 
      "id": "https://www.v2ex.com/t/507194", 
      "date_published": "2018-11-12T15:38:37+00:00", 
      "content_html": "<p>\u6211\u4eec\u6709\u5f88\u591a\u987e\u5ba2\u7684\u804a\u5929\u6587\u5b57\u8bed\u97f3\u5bf9\u8bdd\u3001\u5feb\u5546\u901a\u7b49\u54a8\u8be2\u5e73\u53f0\u7684\u5bf9\u8bdd\u3001\u7535\u8bdd\u56de\u8bbf\u7684\u8bed\u97f3\u3001\u672f\u524d\u672f\u540e\u7167\u7247\u3001360 \u5ea6\u5ba2\u6237\u4fe1\u606f\u3001\u8d2d\u4e70\u6cbb\u7597\u9879\u76ee\u7b49\uff0c\u91cc\u9762\u5c24\u5176\u5ba2\u6237\u7167\u7247\u6d89\u53ca\u9690\u79c1\u5982\u679c\u4e0a\u4f20\u5230\u817e\u8baf\u6216\u767e\u5ea6\u7684\u670d\u52a1\u5668\u4f1a\u4e0d\u5b89\u5168\uff0c\u6709\u63d0\u4f9b\u79bb\u7ebf\u7248\u672c\u4f9b\u4f7f\u7528\u5417\uff1f \u5728\u4fdd\u969c\u6570\u636e\u5b89\u5168\u4e0b\u5176\u4e2d\u53ef\u4f9b\u4e8e\u56fe\u50cf\u8bc6\u522b\u3001\u8bed\u97f3\u8bc6\u522b\u3001\u6587\u5b57\u81ea\u7136\u8bed\u8a00\u5904\u7406\u7684\u6709\u54ea\u4e9b\uff1f\u6280\u672f\u4e0a\u8fbe\u5230\u4e86\uff0c\u5e94\u7528\u573a\u666f\u8fd8\u4e0d\u77e5\u9053\u3002\n\u533b\u52a9\u5728\u8bca\u95f4\u5199\u75c5\u5386\u53ef\u4ee5\u8003\u8651\u7528\u8bed\u97f3\u8f6c\u6587\u5b57,\u4f46\u662f\u533b\u5b66\u672f\u8bed\u4e0d\u77e5\u9053\u73b0\u5728\u7684\u6280\u672f\u80fd\u4e0d\u80fd\u505a\u5230. \u6211\u4eec\u7684\u75c5\u5386\u8fd8\u57fa\u672c\u4e0d\u505a\u79d1\u7814.\u6bd4\u5982\u6211\u4eec\u60f3\u5206\u6790\u987e\u5ba2\u7684\u8138\u7167\u7247\u5f97\u51fa\u4ed6\u662f\u4ec0\u4e48\u773c\u775b\u3001\u4ec0\u4e48\u80a4\u8272\u3001\u4ec0\u4e48\u8138\u578b\u7b49\u8fd9\u4e2a\u6709\u79bb\u7ebf\u7684\u5f00\u6e90\u65b9\u6848\u5417\uff1f</p>\n<p><img src=\"https://user-images.githubusercontent.com/1269898/48357506-bdb45480-e6d3-11e8-855e-695bad60bd43.png\" alt=\"image\"></p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/mindrainbow", 
        "name": "mindrainbow", 
        "avatar": "https://cdn.v2ex.com/avatar/072b/320c/362168_large.png?m=1584935117"
      }, 
      "url": "https://www.v2ex.com/t/506456", 
      "title": "Intel \u51a5\u738b\u5ce1\u8c37 2\uff0c\u8fd9\u4e2a\u673a\u5668\u7684 AMD RX Vega M GH \u82af\u7247\u7ec4\u53ef\u4ee5\u8dd1 TensorFlow \u5417\uff1f", 
      "id": "https://www.v2ex.com/t/506456", 
      "date_published": "2018-11-10T08:06:06+00:00", 
      "content_html": "<p>\u60f3\u4e70\u4e00\u4e2a\u51a5\u738b\u5ce1\u8c37 2 \u4f5c\u4e3a\u5165\u95e8\u7684\u6df1\u5ea6\u5b66\u4e60\u5f00\u53d1\u73af\u5883\uff0c\u5728\u4e0a\u8fb9\u5199\u5199\u4ee3\u7801\uff0c\u4e91\u7aef\u4f5c\u4e3a\u751f\u4ea7\u73af\u5883\u3002\u542c\u8bf4 ROCm \u73b0\u5728\u53ef\u4ee5\u652f\u6301\u8dd1 TensorFlow \u4e86\uff0c\u8fd9\u4e2a\u82af\u7247\u7ec4\u53ef\u4ee5\u8dd1\u5417\uff1f</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/1722332572", 
        "name": "1722332572", 
        "avatar": "https://cdn.v2ex.com/avatar/31e3/2c4d/135473_large.png?m=1739072858"
      }, 
      "url": "https://www.v2ex.com/t/505888", 
      "date_modified": "2018-11-08T11:38:38+00:00", 
      "content_html": "\u78d0\u521b TensorFlow \u7cfb\u5217\u6559\u6750\uff1a <a target=\"_blank\" href=\"http://panchuang.net\" rel=\"nofollow\">http://panchuang.net</a><br /><br />TensorFlow \u7cfb\u5217\u4e13\u9898\uff08\u4e00\uff09\uff1a\u673a\u5668\u5b66\u4e60\u57fa\u7840<br /><a target=\"_blank\" href=\"http://panchuang.net/2018/10/29/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E4%B8%80%EF%BC%89%EF%BC%9A%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E5%9F%BA%E7%A1%80/\" rel=\"nofollow\">http://panchuang.net/2018/10/29/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E4%B8%80%EF%BC%89%EF%BC%9A%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E5%9F%BA%E7%A1%80/</a><br /><br />TensorFlow \u7cfb\u5217\u4e13\u9898\uff08\u4e8c\uff09\uff1a\u673a\u5668\u5b66\u4e60\u57fa\u7840<br /><a target=\"_blank\" href=\"http://panchuang.net/2018/10/26/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E4%BA%8C%EF%BC%89%EF%BC%9A%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E5%9F%BA%E7%A1%80/\" rel=\"nofollow\">http://panchuang.net/2018/10/26/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E4%BA%8C%EF%BC%89%EF%BC%9A%E6%9C%BA%E5%99%A8%E5%AD%A6%E4%B9%A0%E5%9F%BA%E7%A1%80/</a><br /><br />TensorFlow \u7cfb\u5217\u4e13\u9898\uff08\u4e09\uff09\uff1a\u6df1\u5ea6\u5b66\u4e60\u7b80\u4ecb<br /><a target=\"_blank\" href=\"http://panchuang.net/2018/10/29/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E4%B8%89%EF%BC%89%EF%BC%9A%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AE%80%E4%BB%8B/\" rel=\"nofollow\">http://panchuang.net/2018/10/29/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E4%B8%89%EF%BC%89%EF%BC%9A%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E7%AE%80%E4%BB%8B/</a><br /><br />Tensorflow \u7cfb\u5217\u4e13\u9898\uff08\u56db\uff09\uff1a\u795e\u7ecf\u7f51\u7edc\u7bc7\u4e4b\u524d\u9988\u795e\u7ecf\u7f51\u7edc\u7efc\u8ff0<br /><a target=\"_blank\" href=\"http://panchuang.net/2018/11/02/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E5%9B%9B%EF%BC%89%EF%BC%9A%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C%E7%AF%87%E4%B9%8B%E5%89%8D%E9%A6%88%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C%E7%BB%BC/\" rel=\"nofollow\">http://panchuang.net/2018/11/02/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E5%9B%9B%EF%BC%89%EF%BC%9A%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C%E7%AF%87%E4%B9%8B%E5%89%8D%E9%A6%88%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C%E7%BB%BC/</a><br /><br />TensorFlow \u7cfb\u5217\u4e13\u9898\uff08\u4e94\uff09\uff1aBP \u7b97\u6cd5\u539f\u7406<br /><a target=\"_blank\" href=\"http://panchuang.net/2018/11/05/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E4%BA%94%EF%BC%89%EF%BC%9Abp%E7%AE%97%E6%B3%95%E5%8E%9F%E7%90%86/\" rel=\"nofollow\">http://panchuang.net/2018/11/05/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E4%BA%94%EF%BC%89%EF%BC%9Abp%E7%AE%97%E6%B3%95%E5%8E%9F%E7%90%86/</a><br /><br />TensorFlow \u7cfb\u5217\u4e13\u9898\uff08\u516d\uff09\uff1a\u5b9e\u6218\u9879\u76ee Mnist \u624b\u5199\u6570\u636e\u96c6\u8bc6\u522b<br /><a target=\"_blank\" href=\"http://panchuang.net/2018/11/07/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E5%85%AD%EF%BC%89%EF%BC%9A%E5%AE%9E%E6%88%98%E9%A1%B9%E7%9B%AEmnist%E6%89%8B%E5%86%99%E6%95%B0%E6%8D%AE%E9%9B%86%E8%AF%86%E5%88%AB/\" rel=\"nofollow\">http://panchuang.net/2018/11/07/tensorflow%E7%B3%BB%E5%88%97%E4%B8%93%E9%A2%98%EF%BC%88%E5%85%AD%EF%BC%89%EF%BC%9A%E5%AE%9E%E6%88%98%E9%A1%B9%E7%9B%AEmnist%E6%89%8B%E5%86%99%E6%95%B0%E6%8D%AE%E9%9B%86%E8%AF%86%E5%88%AB/</a>", 
      "date_published": "2018-11-08T11:38:15+00:00", 
      "title": "TensorFlow \u7cfb\u5217\u6559\u7a0b\u4e13\u9898", 
      "id": "https://www.v2ex.com/t/505888"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/1722332572", 
        "name": "1722332572", 
        "avatar": "https://cdn.v2ex.com/avatar/31e3/2c4d/135473_large.png?m=1739072858"
      }, 
      "url": "https://www.v2ex.com/t/472748", 
      "title": "TensorFlow 1.9 \u5b98\u65b9\u5165\u95e8\u5b9e\u4f8b\uff1a\u65e0\u9700\u914d\u7f6e\u73af\u5883\uff0c\u5728\u7ebf\u5b66\u4e60 TensorFlow", 
      "id": "https://www.v2ex.com/t/472748", 
      "date_published": "2018-07-20T10:35:10+00:00", 
      "content_html": "<p>TensorFlow \u662f\u4e00\u4e2a\u7528\u4e8e\u7814\u7a76\u548c\u751f\u4ea7\u7684\u5f00\u6e90\u673a\u5668\u5b66\u4e60\u5e93\u3002TensorFlow \u4e3a\u521d\u5b66\u8005\u548c\u4e13\u5bb6\u63d0\u4f9b API\uff0c\u4ee5\u4fbf\u4e3a\u684c\u9762\uff0c\u79fb\u52a8\uff0cWeb \u548c\u4e91\u5f00\u53d1\u3002TensorFlow 1.9 \u5b98\u65b9\u5165\u95e8\u5b9e\u4f8b\uff0c\u6539\u4e3a Keras \u7248\u672c\uff0c\u4ee3\u7801\u6bd4\u8f83\u7b80\u6d01\uff0c\u65e0\u9700\u914d\u7f6e\u73af\u5883\uff0c\u53ef\u4ee5\u5728\u7ebf\u5b66\u4e60 TensorFlow\u3002</p>\n<pre><code># \u5f15\u5165 TensorFlow \u5e93\nimport tensorflow as tf\n# \u5f15\u5165 mnist \u6570\u636e\u96c6\nmnist = tf.keras.datasets.mnist\n# \u8f7d\u5165\u8bad\u7ec3\u96c6\u548c\u6d4b\u8bd5\u96c6\n(x_train, y_train),(x_test, y_test) = mnist.load_data()\nx_train, x_test = x_train / 255.0, x_test / 255.0\n# \u5efa\u7acb Keras \u5e8f\u5217\u6a21\u578b\nmodel = tf.keras.models.Sequential([\n  tf.keras.layers.Flatten(),\n  tf.keras.layers.Dense(512, activation=tf.nn.relu),\n  tf.keras.layers.Dropout(0.2),\n  tf.keras.layers.Dense(10, activation=tf.nn.softmax)\n])\n# \u7f16\u8bd1\u6a21\u578b\uff0coptimizer \u4f18\u5316\u51fd\u6570\uff0closs \u635f\u5931\u51fd\u6570,metrics \u8bc4\u4ef7\u6807\u51c6\nmodel.compile(optimizer='adam',\n              loss='sparse_categorical_crossentropy',\n              metrics=['accuracy'])\n# \u8bad\u7ec3\u6a21\u578b\nmodel.fit(x_train, y_train, epochs=5)\n# \u8bc4\u4f30\u6a21\u578b\nmodel.evaluate(x_test, y_test)\n</code></pre>\n<p>Google \u63d0\u4f9b\u4e86\u5728\u7ebf\u7684 Jupyter \u8fd0\u884c\u73af\u5883\u53ef\u4ee5\u76f4\u63a5\u8fd0\u884c\u4ee5\u4e0a\u5165\u95e8\u5b9e\u4f8b\uff0c\u4e0d\u9700\u8981\u914d\u7f6e\u73af\u5883\u3002</p>\n<img alt=\"\" src=\"http://tf86.com/wp-content/uploads/2018/07/WechatIMG774.jpeg\">\n<p>\u5185\u5bb9\u7ffb\u8bd1\u81ea\uff1a <a href=\"https://www.tensorflow.org/tutorials/\" rel=\"nofollow\">https://www.tensorflow.org/tutorials/</a></p>\n<p>TensorFlow \u6559\u7a0b\uff1a <a href=\"http://tf86.com/2018/07/20/get-started-with-tensorflow/\" rel=\"nofollow\">http://tf86.com/2018/07/20/get-started-with-tensorflow/</a></p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/Hsinyao", 
        "name": "Hsinyao", 
        "avatar": "https://cdn.v2ex.com/avatar/1958/e441/293313_large.png?m=1740301530"
      }, 
      "url": "https://www.v2ex.com/t/471312", 
      "title": "CNN \u8bad\u7ec3\u65f6\uff0c\u4e3a\u4ec0\u4e48\u8981\u5c06\u8f93\u5165\u56fe\u7247\u7684\u50cf\u7d20\u53d6\u503c\u8303\u56f4\u4ece 0~255 \u8f6c\u6362\u4e3a 0~1", 
      "id": "https://www.v2ex.com/t/471312", 
      "date_published": "2018-07-16T06:21:13+00:00", 
      "content_html": "<p>\u5982\u9898\uff0c\u8fd9\u4e2a\u95ee\u9898\u56f0\u6270\u5c0f\u5f1f\u4e00\u4e2a\u4e2d\u5348\u3002\u6211\u628a\u539f\u56fe\u7247(0~255)\u76f4\u63a5\u8f93\u5165\u6a21\u578b\uff0c\u8bad\u7ec3\u7684\u51c6\u786e\u7387\u6536\u655b\u5728 10%\u5de6\u53f3\uff0c\u6309\u7167\u7f51\u4e0a\u7684\u4ee3\u7801\uff0c\u5c06\u6bcf\u4e2a\u50cf\u7d20\u7684\u53d6\u503c\u9664\u4ee5 255\uff0c\u4f7f\u6240\u6709\u50cf\u7d20\u7684\u53d6\u503c\u4ece 0~255 \u8f6c\u5230 0~1\uff0c\u795e\u7ecf\u7f51\u7edc\u7684\u8bad\u7ec3\u6548\u679c\u5c31\u4f1a\u5de8\u5927\u63d0\u5347\uff0c\u8bf7\u95ee\u8fd9\u662f\u4ec0\u4e48\u539f\u7406\uff1f\u4e3a\u4ec0\u4e48\u8981\u8f6c\u6362\u4e00\u4e0b\uff1f</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/eccstartup", 
        "name": "eccstartup", 
        "avatar": "https://cdn.v2ex.com/avatar/5d76/7a83/66546_large.png?m=1486289910"
      }, 
      "url": "https://www.v2ex.com/t/465381", 
      "date_modified": "2020-07-14T17:27:09+00:00", 
      "content_html": "\u6709\u6ca1\u6709\u4ec0\u4e48\u53ef\u4ee5\u5feb\u901f\u5165\u95e8\u7684\u6750\u6599\uff0c\u770b\u4ee3\u7801\u80fd\u770b\u61c2\uff0c\u4f46\u662f\u4e0d\u77e5\u9053\u4e3a\u5565\u4f1a\u60f3\u5230\u8fd9\u4e9b\u51fd\u6570\u3002", 
      "date_published": "2018-06-24T02:26:08+00:00", 
      "title": "tensorflow \u7684 API \u600e\u4e48\u80fd\u591f\u5feb\u901f\u638c\u63e1\u5462\uff1f", 
      "id": "https://www.v2ex.com/t/465381"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/WildCat", 
        "name": "WildCat", 
        "avatar": "https://cdn.v2ex.com/avatar/baf2/9ae7/39784_large.png?m=1710948175"
      }, 
      "url": "https://www.v2ex.com/t/463923", 
      "date_modified": "2018-06-18T12:30:41+00:00", 
      "content_html": "<a target=\"_blank\" href=\"http://www.rethink.fun/index.php/2018/05/03/hand_writing_recgnition/\" rel=\"nofollow\">http://www.rethink.fun/index.php/2018/05/03/hand_writing_recgnition/</a>\r<br /><a target=\"_blank\" href=\"http://www.rethink.fun/index.php/2018/05/04/estimator/\" rel=\"nofollow\">http://www.rethink.fun/index.php/2018/05/04/estimator/</a>\r<br /><a target=\"_blank\" href=\"http://www.rethink.fun/index.php/2018/05/05/recognize3/\" rel=\"nofollow\">http://www.rethink.fun/index.php/2018/05/05/recognize3/</a>", 
      "date_published": "2018-06-18T12:27:43+00:00", 
      "title": "\u7528 TensorFlow \u5b9e\u73b0\u4e00\u4e2a\u624b\u5199\u6c49\u5b57\u8bc6\u522b", 
      "id": "https://www.v2ex.com/t/463923"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/4thmagi", 
        "name": "4thmagi", 
        "avatar": "https://cdn.v2ex.com/avatar/3bdc/c8dc/244937_large.png?m=1554689534"
      }, 
      "url": "https://www.v2ex.com/t/462623", 
      "date_modified": "2018-06-12T15:24:06+00:00", 
      "content_html": "\u6211\u5199\u4e86\u4e00\u4e2a\u7f51\u9875\uff0c\u662f\u57fa\u4e8e django \u7684\u3002\u8be5\u7f51\u9875\u901a\u8fc7 session \u5b9e\u73b0\u4e86\u7528\u6237\u767b\u5f55\u72b6\u6001\u7684\u5224\u65ad\u3002\u7528\u6237\u5728\u767b\u5f55\u60c5\u51b5\u4e0b\u53ef\u4ee5\u53d1\u5e03\u4fe1\u606f\uff0c\u5176\u4e2d\u53ef\u4ee5\u5305\u542b\u56fe\u7247\u3002\u7136\u540e\u53ef\u4ee5\u5bf9\u53d1\u5e03\u7684\u56fe\u7247\u8fdb\u884c\u6570\u5b57\u7684\u8bc6\u522b\u3002\r<br />\u7f51\u9875\u540e\u7aef\u6211\u4f7f\u7528\u7684\u662f django \u6846\u67b6,\u4e0a\u4f20\u4e00\u5f20\u56fe\u7247\u4f20\u5165\u57fa\u4e8e tensorflow \u7684 keras \u6a21\u578b\u8fdb\u884c\u9884\u6d4b,\u91cd\u590d\u9884\u6d4b\u65f6,\u62a5\u544a\u9519\u8bef\uff1aValueError: Tensor Tensor(&quot;Placeholder:0&quot;, shape=(3, 3, 1, 32), dtype=float32)\u3002\u67e5\u4e86\u4e00\u4e0b\uff0c\u539f\u56e0\u5927\u6982\u662f\u7b2c\u4e8c\u6b21\u9884\u6d4b\u65f6,model \u5e95\u5c42 tensorflow \u7684 session \u4e2d\u8fd8\u6709\u6570\u636e\u3002\r<br />\u5728\u7f51\u4e0a\u627e\u5230\u7684\u89e3\u51b3\u65b9\u6cd5\u662f\u8c03\u7528\u6a21\u578b\u524d\u9762\u52a0\u4e00\u53e5 keras.backend.clear_session()\u3002\u4f46\u662f\u8fd9\u6837\u7684\u8bdd\u5c31\u628a\u6211\u7684\u767b\u5f55\u72b6\u6001\u4e5f\u6e05\u7a7a\u4e86\uff0c\u7f51\u9875\u76f4\u63a5\u5d29\u6e83\u4e86\u3002\u63d0\u793a python \u5df2\u505c\u6b62\u5de5\u4f5c\u3002\r<br />\u8bf7\u95ee\u6709\u4ec0\u4e48\u529e\u6cd5\u53ef\u4ee5\u53ea\u6e05\u7a7a session \u4e2d\u7684 TensorFlow \u7684\u76f8\u5173\u6570\u636e\uff1f", 
      "date_published": "2018-06-12T15:22:32+00:00", 
      "title": "keras \u6846\u67b6\u4e0e django \u6846\u67b6\u4e00\u8d77\u4f7f\u7528 \u53cd\u590d\u8c03\u7528 model \u6a21\u578b \u51fa\u9519", 
      "id": "https://www.v2ex.com/t/462623"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/helloworld12", 
        "name": "helloworld12", 
        "avatar": "https://cdn.v2ex.com/gravatar/422d620917d5f54ce15bf728df86b5db?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/456206", 
      "date_modified": "2018-05-19T16:47:26+00:00", 
      "content_html": "<p>\u4ee3\u7801\u5982\u4e0b:</p>\n<pre><code>#encoding: utf-8\nimport tensorflow as tf\nfrom numpy.random import RandomState\n\n\n# \u83b7\u53d6\u4e00\u5c42\u795e\u7ecf\u7f51\u7edc\u8fb9\u4e0a\u7684\u6743\u91cd\uff0c\u5e76\u5c06\u8fd9\u4e2a\u6743\u91cd\u7684 L2 \u6b63\u5219\u5316\u635f\u5931\u52a0\u5165\u540d\u79f0\u4e3a'losses'\u7684\u96c6\u5408\u4e2d\ndef get_weight(shape, lambdaF):\n    var = tf.Variable(tf.random_normal(shape))\n    # add_to_collection \u51fd\u6570\u5c06\u8fd9\u4e2a\u65b0\u751f\u6210\u53d8\u91cf\u7684 L2 \u6b63\u5219\u5316\u635f\u5931\u52a0\u5165\u96c6\u5408\u3002\n    # \u8fd9\u4e2a\u51fd\u6570\u7684\u7b2c\u4e00\u4e2a\u53c2\u6570'losses'\uff0c \u662f\u96c6\u5408\u7684\u540d\u5b57\uff0c\u7b2c\u4e8c\u4e2a\u53c2\u6570\u662f\u8981\u52a0\u5165\u8fd9\u4e2a\u96c6\u5408\u7684\u5185\u5bb9\n    tf.add_to_collection('losses', tf.contrib.layers.l2_regularizer(lambdaF)(var))\n\n    return var\n\nx = tf.placeholder(tf.float32, shape=(None, 2), name='x-input')\ny_ = tf.placeholder(tf.float32, shape=(None, 1), name='y-input')\nbatch_size = 8\n\n# \u5b9a\u4e49\u4e86\u6bcf\u4e00\u5c42\u7f51\u7edc\u4e2d\u8282\u70b9\u7684\u4e2a\u6570\nlayer_dimension = [2, 10, 10, 10, 1]\n# \u795e\u7ecf\u7f51\u7edc\u7684\u5c42\u6570\u3002\nn_layers = len(layer_dimension)\n\n# \u8fd9\u4e2a\u53d8\u91cf\u7ef4\u62a4\u524d\u5411\u4f20\u64ad\u65f6\u6700\u6df1\u5c42\u7684\u8282\u70b9\uff0c\u5f00\u59cb\u7684\u65f6\u5019\u5c31\u662f\u8f93\u5165\u5c42\u3002\ncur_layer = x\n# \u5f53\u524d\u5c42\u7684\u8282\u70b9\u4e2a\u6570\nin_dimension = layer_dimension[0]\n\n# \u901a\u8fc7\u4e00\u4e2a\u5faa\u73af\u6765\u751f\u6210 5 \u5c42\u5168\u8fde\u63a5\u7684\u795e\u7ecf\u7f51\u7edc\u7ed3\u6784\nfor i in range(1, n_layers):\n    # layer_dimension[i]\u4e3a\u4e0b\u4e00\u5c42\u7684\u8282\u70b9\u4e2a\u6570\u3002\n    out_dimension = layer_dimension[i]\n    # \u751f\u6210\u5f53\u524d\u5c42\u4e2d\u6743\u91cd\u7684\u53d8\u91cf\uff0c\u5e76\u5c06\u8fd9\u4e2a\u53d8\u91cf\u7684 L2 \u6b63\u5219\u5316\u635f\u5931\u52a0\u5165\u8ba1\u7b97\u56fe\u4e0a\u7684\u96c6\u5408\u3002\n    weight = get_weight([in_dimension, out_dimension], 0.001)\n    bias = tf.Variable(tf.constant(0.1, shape=[out_dimension]))\n    # \u4f7f\u7528 ReLU \u6fc0\u6d3b\u51fd\u6570\n    cur_layer = tf.nn.relu(tf.matmul(cur_layer, weight) + bias)\n    # \u8fdb\u5165\u4e0b\u4e00\u5c42\u4e4b\u524d\u5c06\u4e0b\u4e00\u5c42\u7684\u8282\u70b9\u6570\u66f4\u65b0\u4e3a\u5f53\u524d\u5c42\u8282\u70b9\u4e2a\u6570\n    in_dimension = layer_dimension[i]\n\n# \u5728\u5b9a\u4e49\u795e\u7ecf\u7f51\u7edc\u524d\u5411\u4f20\u64ad\u7684\u540c\u65f6\u5df2\u7ecf\u5c06\u6240\u6709\u7684 L2 \u6b63\u5219\u5316\u635f\u5931\u52a0\u5165\u56fe\u4e0a\u7684\u96c6\u5408\uff0c\n# \u8fd9\u91cc\u53ea\u9700\u8981\u8ba1\u7b97\u523b\u753b\u6a21\u578b\u5728\u8bad\u7ec3\u6570\u636e\u4e0a\u8868\u73b0\u7684\u635f\u5931\u51fd\u6570\nmse_loss = tf.reduce_mean(tf.square(y_ - cur_layer))\n\n# \u5c06\u5747\u65b9\u8bef\u5dee\u635f\u5931\u51fd\u6570\u52a0\u5165\u635f\u5931\u96c6\u5408\ntf.add_to_collection('losses', mse_loss)\n\n# get_collection \u8fd4\u56de\u4e00\u4e2a\u5217\u8868\uff0c \u8fd9\u4e2a\u5217\u8868\u662f\u6240\u6709\u8fd9\u4e2a\u96c6\u5408\u7684\u5143\u7d20\u3002\u5728\u8fd9\u4e2a\u6837\u4f8b\u4e2d\uff0c\n# \u8fd9\u4e9b\u5143\u7d20\u5c31\u662f\u635f\u5931\u51fd\u6570\u7684\u4e0d\u540c\u90e8\u5206\uff0c\u5c06\u5b83\u4eec\u52a0\u8d77\u6765\u5c31\u53ef\u4ee5\u5f97\u5230\u6700\u7ec8\u7684\u635f\u5931\u51fd\u6570\nloss = tf.add_n(tf.get_collection('losses'))\ntran = tf.train.AdamOptimizer().minimize(loss)\n\nrdm = RandomState(1)\ndataset_size = 12800\nX = rdm.rand(dataset_size, 2)\nY = [ (x1*x2 +  rdm.rand()/10.0 - 0.05,) for (x1, x2) in X]\n\n\nwith tf.Session() as sess:\n    sess.run(tf.global_variables_initializer())\n\n    STEPS = 1500000\n    for i in range(STEPS):\n        start = (i * batch_size) % dataset_size\n        end = min(dataset_size, start+batch_size)\n\n        sess.run(tran, feed_dict={x:X[start:end], y_:Y[start:end]})\n        if i % 200 == 0:\n            print(sess.run(loss, feed_dict={x:X, y_:Y}))\n</code></pre>\n<p>\u4e3b\u8981\u7591\u95ee\u662f:</p>\n<pre><code>X = rdm.rand(dataset_size, 2)\nY = [ (x1*x2 +  rdm.rand()/10.0 - 0.05,) for (x1, x2) in X]\n</code></pre>\n<p>\u4e3a\u4ec0\u4e48,\u90a3\u4e9b\u6743\u91cd\u53c2\u6570,\u80fd\u62df\u5408\u51fa X \u548c Y \u7684\u6620\u5c04\u5173\u7cfb,Y \u662f\u7b49\u4e8e x1*x2,\u8fd9\u4e00\u70b9\u4e5f\u4e0d\u7ebf\u6027\u554a<br>\n\u8fd0\u884c\u4e86\u4e0b,loss \u53ef\u4ee5\u5c0f\u5230 0.0035882844</p>\n", 
      "date_published": "2018-05-19T14:37:44+00:00", 
      "title": "TensorFlow \u521d\u5b66\u8005,\u8bf7\u6559\u4e0b", 
      "id": "https://www.v2ex.com/t/456206"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/Asice", 
        "name": "Asice", 
        "avatar": "https://cdn.v2ex.com/gravatar/299caae2dedb90994785332853b19c52?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/454530", 
      "title": "Tensorflow \u5c0f\u8bf4\u53d8\u52a8\u753b", 
      "id": "https://www.v2ex.com/t/454530", 
      "date_published": "2018-05-13T12:30:53+00:00", 
      "content_html": "<p>\u300a\u5b89\u5a1c\u5361\u5217\u5c3c\u5a1c\u300b\u6587\u672c\u751f\u6210\u2014\u2014\u5229\u7528 TensorFlow \u6784\u5efa LSTM \u6a21\u578b\n<a href=\"https://zhuanlan.zhihu.com/p/27087310\" rel=\"nofollow\">https://zhuanlan.zhihu.com/p/27087310</a>\n\u4eb2\u6d4b\u597d\u7528\uff01 Google \u53d1\u5e03\u4e86\u4e00\u4e2a\u65b0\u7684 Tensorflow \u7269\u4f53\u8bc6\u522b API\n<a href=\"http://36kr.com/p/5090812.html\" rel=\"nofollow\">http://36kr.com/p/5090812.html</a>\n\u6309\u8fd9 2 \u7bc7\u6587\u7ae0\uff0c\u5c0f\u8bf4\u53d8\u6f2b\u753b\u662f\u53ef\u4ee5\u5b9e\u73b0\u7684\uff0c\u5c0f\u8bf4\u53d8\u52a8\u753b\u5e94\u8be5\u4e5f\u662f\u53ef\u4ee5\u5b9e\u73b0\u7684\u3002\n\u5c31\u662f\u6a21\u578b\u7684\u8bad\u7ec3\u91cf\u53ef\u80fd\u8981\u975e\u5e38\u5927</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/smy20011", 
        "name": "smy20011", 
        "avatar": "https://cdn.v2ex.com/gravatar/66b2b2829bf7e92105f5f949138d8490?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/433896", 
      "date_modified": "2018-03-01T06:29:20+00:00", 
      "content_html": "<p>\u8fd9\u4e2a\u9879\u76ee\u6211\u4e5f\u5e2e\u5fd9\u4e86\u8fdb\u884c\u4e86\u4e00\u4e9b\u7ffb\u8bd1\uff0c\u5fd9\u4e86\u4e00\u5e74\u4ed6\u4eec\u603b\u7b97\u6709\u70b9\u7ed3\u679c\u4e86\u3002\u8fd9\u4e2a\u662f\u6211\u4eec\u5185\u90e8\u7528\u7684\u673a\u5668\u5b66\u4e60\u5165\u95e8\u6559\u7a0b\uff0c\u73b0\u5728\u4ed6\u4eec\u5f00\u653e\u7ed9\u5927\u4f17\u4f7f\u7528\u4e86\u3002</p>\n<p>\u5899\u5185\u7684\u5730\u5740\uff1a\n<a href=\"https://developers.google.cn/machine-learning/crash-course/\" rel=\"nofollow\">https://developers.google.cn/machine-learning/crash-course/</a></p>\n<p>\u5982\u679c\u80fd\u7ffb\u5899\u7684\u8bdd\uff0c\u6211\u8fd8\u662f\u63a8\u8350\u539f\u7248\u7684\uff0c\u56e0\u4e3a\u56fd\u5185\u4e0d\u80fd youtube\uff0c\u4ed6\u4eec\u7528\u7684 tts \u6765\u5408\u6210\u4e2d\u6587\u7684\u8bed\u97f3\u3002.com \u7248\u672c\u7684\u7528\u7684\u662f youtube \u6765\u64ad\u653e\u89c6\u9891\uff0c\u542c\u7684\u76f8\u5bf9\u8212\u670d\u4e00\u70b9\u3002\n<a href=\"https://developers.google.com/machine-learning/crash-course/\" rel=\"nofollow\">https://developers.google.com/machine-learning/crash-course/</a></p>\n<p>\u53e6\u5916\uff0cAD Blocker \u4f1a\u8ba9\u89c6\u9891\u653e\u4e0d\u51fa\u6765\uff0c\u4e0a\u8fd9\u4e2a\u7f51\u7ad9\u7684\u65f6\u5019\u5efa\u8bae\u628a adblocker \u5173\u6389\u3002\n\u6709\u5565\u610f\u89c1\u53ef\u4ee5\u63d0\u51fa\u6765\uff0c\u6211\u770b\u770b\u80fd\u4e0d\u80fd\u8f6c\u7ed9\u505a\u8fd9\u4e2a\u7684\u4eba\u3002</p>\n<p>\u4ee5\u4e0a</p>\n", 
      "date_published": "2018-03-01T06:27:52+00:00", 
      "title": "\u5b89\u5229\u4e0b Google \u6700\u8fd1\u51fa\u7684\u673a\u5668\u5b66\u4e60\u6559\u7a0b", 
      "id": "https://www.v2ex.com/t/433896"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/miniliuke", 
        "name": "miniliuke", 
        "avatar": "https://cdn.v2ex.com/gravatar/d1f6edc591d0494de75467adc42aa4f5?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/430058", 
      "title": "tensorflow \u4fdd\u5b58\u597d\u7684\u6a21\u578b\u53c2\u6570\u5728\u4e0d\u540c\u7248\u672c\u4e0b", 
      "id": "https://www.v2ex.com/t/430058", 
      "date_published": "2018-02-10T12:37:10+00:00", 
      "content_html": "<p>tensorflow \u4fdd\u5b58\u597d\u7684\u6a21\u578b\u53c2\u6570\uff0c\u7248\u672c\u662f 1.2......\u6211\u7684\u7248\u672c\u662f 1.4......\u4f1a\u6709 Warning \u4f46\u80fd\u7528,\u6211\u60f3\u628a\u4ed6\u4fdd\u5b58\u4e3a 1.4 \u7248\u672c\u4e0b\u7684......\u5c31\u662f\u52a0\u8f7d\u56fe\u548c\u53c2\u6570\u7136\u540e\u4fdd\u5b58\uff0c\u7ed3\u679c\u4e00\u76f4\u62a5\u9519......\u600e\u4e48\u529e\uff1f\u96be\u9053\u53ea\u80fd\u91cd\u65b0\u6784\u9020 model......</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/yangyaofei", 
        "name": "yangyaofei", 
        "avatar": "https://cdn.v2ex.com/gravatar/9247611e1af0fa4e3eeeeb5f1b8bc79d?s=73&d=retro"
      }, 
      "url": "https://www.v2ex.com/t/429720", 
      "title": "\u8bf7\u95ee,\u5982\u4f55\u5b9e\u73b0 4D \u7684\u5377\u79ef", 
      "id": "https://www.v2ex.com/t/429720", 
      "date_published": "2018-02-09T05:14:30+00:00", 
      "content_html": "<p>\u6700\u8fd1\u6709\u4e2a\u8111\u6d1e,\u60f3\u5b9e\u73b0\u4e00\u4e0b,\u4f46\u662f\u8981\u7528\u5230 4 \u7ef4\u4f5c\u4e3a\u8f93\u5165,\u4f46\u662f\u770b\u5230 TensorFlow \u7b49\u6846\u67b6,\u6ca1\u6709\u80fd 4d \u7684\u5377\u79ef,\u8bf7\u95ee,\u6709\u4ec0\u4e48\u65b9\u6cd5\u53ef\u4ee5\u5b9e\u73b0.</p>\n<p>\u53e6\u5916,\u5982\u679c\u81ea\u5df1\u5168\u624b\u52a8\u5b9e\u73b0\u4f1a\u4e0d\u4f1a\u5f88\u8017\u65f6\u95f4,\u6bd5\u7adf\u4e0d\u662f\u4e00\u4e2a\u5f88\u9760\u8c31\u7684\u8111\u6d1e...</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/panins", 
        "name": "panins", 
        "avatar": "https://cdn.v2ex.com/avatar/3ea2/b672/123681_large.png?m=1516545522"
      }, 
      "url": "https://www.v2ex.com/t/424758", 
      "title": "expand_dims \u4e2d\u7684\u56f0\u60d1", 
      "id": "https://www.v2ex.com/t/424758", 
      "date_published": "2018-01-21T14:35:24+00:00", 
      "content_html": "<p>\u4e70\u4e86\u4e00\u300a tensorflow \u673a\u5668\u5b66\u4e60\u5b9e\u6218\u6307\u5357\u300b\uff0c\u81ea\u5df1\u5b66\u5230\u4e86\u7b2c\u516d\u7ae0\uff0c\u795e\u7ecf\u7f51\u7edc\u7b97\u6cd5\u3002</p>\n<p>\u4f46\u662f\u6709\u4e2a\u7b80\u5355\u4f8b\u5b50\u5c0f\u5f1f\u4e00\u65f6\u534a\u4f1a\u5361\u5728\u8fd9\u91cc\uff0c\u5927\u5bb6\u662f\u5426\u53ef\u4ee5\u62bd\u7a7a\u6307\u5bfc\u4e00\u4e0b\u3002</p>\n<p>\u8c22\u8c22</p>\n<h3>6.5 \u7528 TENSORFLOW \u5b9e\u73b0\u795e\u7ecf\u7f51\u7edc\u5e38\u89c1\u5c42</h3>\n<p>\u6211\u7684\u95ee\u9898\u662f\u5b9a\u4e49\u5377\u79ef\u5c42\u51fd\u6570\u7684\u65f6\u5019\uff0c\u4e66\u4e0a\u8bf4\u628a 1D \u6269\u5c55\u5230 4D\u3002</p>\n<p>\u4e3a\u4ec0\u4e48 input_4d = tf.expand_dims(input_3d, 3)\u7b2c\u4e8c\u4e2a\u53c2\u6570\u7528\u4e86 3\uff0c\u800c input_1-3d \u90fd\u662f 0</p>\n<pre><code>import tensorflow as tf\nimport numpy as np\nsess = tf.Session()\ndata_size = 25\ndata_1d = np.random.normal(size=data_size)\nx_input_1d = tf.placeholder(dtype=tf.float32, shape=[data_size])\n\n\ndef conv_layer_1d(input_1d, my_filter):\n    # make 1d input into 4d\n    input_2d = tf.expand_dims(input_1d, 0)\n    input_3d = tf.expand_dims(input_2d, 0)\n    input_4d = tf.expand_dims(input_3d, 3)\n    # perform convolution\n    convolution_output = tf.nn.conv2d(input_4d, filter=my_filter,\n                                      strides=[1,1,1,1], padding=\"VALID\")\n    conv_output_1d = tf.squeeze(convolution_output)\n    return(conv_output_1d)\n\nmy_filter = tf.Variable(tf.random_normal(shape=[1,5,1,1]))\nmy_convolution_output = conv_layer_1d(x_input_1d, my_filter)\n......\n</code></pre>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/xwhxbg", 
        "name": "xwhxbg", 
        "avatar": "https://cdn.v2ex.com/avatar/7855/4251/219201_large.png?m=1767925757"
      }, 
      "url": "https://www.v2ex.com/t/421323", 
      "title": "\u4e00\u4e2a\u4e25\u8083\u7684\u5173\u4e8e\u6df1\u5ea6\u5b66\u4e60\u9274\u9ec4\u7684\u95ee\u9898", 
      "id": "https://www.v2ex.com/t/421323", 
      "date_published": "2018-01-09T04:14:26+00:00", 
      "content_html": "<p>\u600e\u4e48\u9274\u522b\u5e73\u80f8\u59b9\u5b50\uff1f\u6211\u8bd5\u56fe\u628a\u5e73\u80f8\u59b9\u5b50\u6807\u8bb0\u4e3a\u9ec4\u56fe\uff0c\u4f46\u662f\u4f1a\u8bef\u4f24\u6ca1\u7a7f\u4e0a\u8863\u7684\u6c49\u5b50\u554a\uff0c\u600e\u4e48\u529e\u5462\uff1f</p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/1722332572", 
        "name": "1722332572", 
        "avatar": "https://cdn.v2ex.com/avatar/31e3/2c4d/135473_large.png?m=1739072858"
      }, 
      "url": "https://www.v2ex.com/t/409724", 
      "title": "\u6211\u5728 B \u7ad9\u5b66\u673a\u5668\u5b66\u4e60\uff08Machine Learning\uff09- \u5434\u6069\u8fbe\uff08Andrew Ng\uff09 [\u4e2d\u82f1\u53cc\u8bed]", 
      "id": "https://www.v2ex.com/t/409724", 
      "date_published": "2017-11-26T13:44:48+00:00", 
      "content_html": "<p>\u6211\u5728 B \u7ad9\u5b66\u673a\u5668\u5b66\u4e60\uff08 Machine Learning \uff09- \u5434\u6069\u8fbe\uff08 Andrew Ng \uff09 [\u4e2d\u82f1\u53cc\u8bed]</p>\n<img alt=\"\" src=\"http://www.tensorflownews.com/wp-content/uploads/2017/11/1511694058247.jpg\">\n<img alt=\"\" src=\"http://www.tensorflownews.com/wp-content/uploads/2017/11/1511694101860.jpg\">\n<img alt=\"\" src=\"http://www.tensorflownews.com/wp-content/uploads/2017/11/WechatIMG39.jpeg\">\n<p>\u89c6\u9891\u5730\u5740\uff1a <a href=\"https://www.bilibili.com/video/av9912938/\" rel=\"nofollow\">https://www.bilibili.com/video/av9912938/</a></p>\n<p>\u66f4\u591a\u6df1\u5ea6\u5b66\u4e60\u8d44\u6e90\uff1a <a href=\"http://tensorflow123.com\" rel=\"nofollow\">http://tensorflow123.com</a></p>\n"
    }, 
    {
      "author": {
        "url": "https://www.v2ex.com/member/1722332572", 
        "name": "1722332572", 
        "avatar": "https://cdn.v2ex.com/avatar/31e3/2c4d/135473_large.png?m=1739072858"
      }, 
      "url": "https://www.v2ex.com/t/409686", 
      "title": "\u6211\u5728 B \u7ad9\u5b66\u4e60\u6df1\u5ea6\u5b66\u4e60\uff08\u751f\u52a8\u5f62\u8c61\uff0c\u8dc3\u7136\u7eb8\u4e0a\uff09", 
      "id": "https://www.v2ex.com/t/409686", 
      "date_published": "2017-11-26T10:58:18+00:00", 
      "content_html": "<p>\u6211\u5728 B \u7ad9\u5b66\u4e60\u6df1\u5ea6\u5b66\u4e60\uff08\u751f\u52a8\u5f62\u8c61\uff0c\u8dc3\u7136\u7eb8\u4e0a\uff09</p>\n<img alt=\"\" src=\"http://www.tensorflownews.com/wp-content/uploads/2017/11/1511693448025.jpg\">\n<img alt=\"\" src=\"http://www.tensorflownews.com/wp-content/uploads/2017/11/1511693406697.jpg\">\n<img alt=\"\" src=\"http://www.tensorflownews.com/wp-content/uploads/2017/11/1511693255737.jpg\">\n<p>\u89c6\u9891\u5730\u5740\uff1a <a href=\"https://www.bilibili.com/video/av16577449/\" rel=\"nofollow\">https://www.bilibili.com/video/av16577449/</a></p>\n<p>\u66f4\u591a\u6df1\u5ea6\u5b66\u4e60\u8d44\u6e90\uff1a <a href=\"http://tensorflow123.com/forum.php?mod=viewthread&amp;tid=99\" rel=\"nofollow\">http://tensorflow123.com/forum.php?mod=viewthread&amp;tid=99</a></p>\n"
    }
  ]
}