{"id":28553,"date":"2025-03-13T23:58:31","date_gmt":"2025-03-13T15:58:31","guid":{"rendered":"https:\/\/www.aisharenet.com\/?p=28553"},"modified":"2025-03-13T23:58:52","modified_gmt":"2025-03-13T15:58:52","slug":"ollama-zai-langchain","status":"publish","type":"post","link":"https:\/\/www.kdjingpai.com\/en\/ollama-zai-langchain\/","title":{"rendered":"Ollama \u5728 LangChain \u4e2d\u7684\u4f7f\u7528 &#8211; Python \u96c6\u6210"},"content":{"rendered":"<h2>\u7b80\u4ecb<\/h2>\n<p>\u672c\u6587\u6863\u4ecb\u7ecd\u4e86\u5982\u4f55\u5728 Python \u73af\u5883\u4e2d\u4f7f\u7528 <a href=\"https:\/\/www.kdjingpai.com\/en\/ollama\/\">Ollama<\/a> \u4e0e LangChain \u96c6\u6210\uff0c\u4ee5\u521b\u5efa\u5f3a\u5927\u7684 AI \u5e94\u7528\u3002Ollama \u662f\u4e00\u4e2a\u5f00\u6e90\u7684\u5927\u8bed\u8a00\u6a21\u578b\u90e8\u7f72\u5de5\u5177\uff0c\u800c LangChain \u5219\u662f\u4e00\u4e2a\u7528\u4e8e\u6784\u5efa\u57fa\u4e8e\u8bed\u8a00\u6a21\u578b\u7684\u5e94\u7528\u7684\u6846\u67b6\u3002\u901a\u8fc7\u7ed3\u5408\u8fd9\u4e24\u8005\uff0c\u6211\u4eec\u53ef\u4ee5\u5728\u672c\u5730\u73af\u5883\u4e2d\u5feb\u901f\u90e8\u7f72\u548c\u4f7f\u7528\u5148\u8fdb\u7684AI\u6a21\u578b\u3002<\/p>\n<blockquote><p>\u6ce8: \u672c\u6587\u6863\u5305\u542b\u6838\u5fc3\u4ee3\u7801\u7247\u6bb5\u548c\u8be6\u7ec6\u89e3\u91ca\u3002\u5b8c\u6574\u4ee3\u7801\u53ef\u5728<a href=\"https:\/\/github.com\/datawhalechina\/handy-ollama\/blob\/main\/notebook\/C5\/ollama_langchain_python.ipynb\">\u6b64Jupyter notebook<\/a>\u4e2d\u627e\u5230\u3002<\/p><\/blockquote>\n<p>&nbsp;<\/p>\n<h2>1. \u73af\u5883\u8bbe\u7f6e<\/h2>\n<h3>\u914d\u7f6e Conda \u73af\u5883<\/h3>\n<p>\u9996\u5148\uff0c\u6211\u4eec\u9700\u8981\u5728 Jupyter \u4e2d\u4f7f\u7528 Conda \u73af\u5883\u3002\u5728\u547d\u4ee4\u884c\u4e2d\u6267\u884c\u4ee5\u4e0b\u547d\u4ee4\uff1a<\/p>\n<pre><code>conda create -n handlm python=3.10 -y\r\nconda activate handlm\r\npip install jupyter\r\npython -m ipykernel install --user --name=handlm\r\n<\/code><\/pre>\n<p>\u6267\u884c\u5b8c\u6bd5\u540e\uff0c\u91cd\u542f Jupyter\uff0c\u5e76\u9009\u62e9\u8be5\u73af\u5883\u7684 Kernel\uff0c\u5982\u56fe\u6240\u793a\uff1a<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-28554\" title=\"Ollama \u5728 LangChain \u4e2d\u7684\u4f7f\u7528 - Python \u96c6\u6210-1\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2526c68628b8077.png\" alt=\"Ollama \u5728 LangChain \u4e2d\u7684\u4f7f\u7528 - Python \u96c6\u6210-1\" width=\"1792\" height=\"196\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2526c68628b8077.png 1792w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2526c68628b8077-768x84.png 768w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2526c68628b8077-1536x168.png 1536w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/2526c68628b8077-18x2.png 18w\" sizes=\"auto, (max-width: 1792px) 100vw, 1792px\" \/><\/p>\n<h3>\u26a0\ufe0f \u6ce8\u610f<\/h3>\n<blockquote><p><strong>\u6ce8\u610f\uff1a<\/strong>\u00a0\u4e5f\u53ef\u4ee5\u4e0d\u4f7f\u7528conda\u865a\u62df\u73af\u5883\uff0c\u76f4\u63a5\u4f7f\u7528\u5168\u5c40\u73af\u5883\u3002<\/p><\/blockquote>\n<h3>\u5b89\u88c5\u4f9d\u8d56<\/h3>\n<p>\u5728\u5f00\u59cb\u4e4b\u524d\uff0c\u6211\u4eec\u9700\u8981\u5b89\u88c5\u4ee5\u4e0b\u5305\uff1a<\/p>\n<ul>\n<li><code>langchain-ollama<\/code>: \u7528\u4e8e\u96c6\u6210 Ollama \u6a21\u578b\u5230 LangChain \u6846\u67b6\u4e2d<\/li>\n<li><code>langchain<\/code>: LangChain \u7684\u6838\u5fc3\u5e93\uff0c\u63d0\u4f9b\u4e86\u6784\u5efa AI \u5e94\u7528\u7684\u5de5\u5177\u548c\u62bd\u8c61<\/li>\n<li><code>langchain-community<\/code>: \u5305\u542b\u4e86\u793e\u533a\u8d21\u732e\u7684\u5404\u79cd\u96c6\u6210\u548c\u5de5\u5177<\/li>\n<li><code>Pillow<\/code>: \u7528\u4e8e\u56fe\u50cf\u5904\u7406\uff0c\u5728\u591a\u6a21\u6001\u4efb\u52a1\u4e2d\u4f1a\u7528\u5230<\/li>\n<li><code>faiss-cpu<\/code>: \u7528\u4e8e\u6784\u5efa\u7b80\u5355 <a href=\"https:\/\/www.kdjingpai.com\/en\/rag\/\">RAG<\/a> \u68c0\u7d22\u5668<\/li>\n<\/ul>\n<p>\u53ef\u4ee5\u901a\u8fc7\u4ee5\u4e0b\u547d\u4ee4\u5b89\u88c5\uff1a<\/p>\n<pre><code>pip install langchain-ollama langchain langchain-community Pillow faiss-cpu\r\n<\/code><\/pre>\n<p>&nbsp;<\/p>\n<h2>2. \u4e0b\u8f7d\u6240\u9700\u6a21\u578b\u5e76\u521d\u59cb\u5316 OllamaLLM<\/h2>\n<h3>\u4e0b\u8f7d llama3.1 \u6a21\u578b<\/h3>\n<ol>\n<li>\u8fdb\u5165\u5b98\u7f51\u00a0https:\/\/ollama.com\/download\u00a0\u4e0b\u8f7d\u5e76\u5b89\u88c5 Ollama \u5230\u53ef\u7528\u7684\u53d7\u652f\u6301\u5e73\u53f0\u3002<\/li>\n<li>\u67e5\u770b\u00a0https:\/\/ollama.ai\/library\u00a0\u4e86\u89e3\u6240\u6709\u53ef\u7528\u7684\u6a21\u578b\u3002<\/li>\n<li>\u901a\u8fc7\u00a0<code>ollama pull &lt;name-of-model&gt;<\/code>\u00a0\u547d\u4ee4\u83b7\u53d6\u53ef\u7528 LLM \u6a21\u578b\uff08\u4f8b\u5982\uff1a<code>ollama pull llama3.1<\/code>\uff09\u3002<\/li>\n<\/ol>\n<p>\u547d\u4ee4\u884c\u6267\u884c\u5b8c\u6bd5\u540e\u5982\u56fe\u6240\u793a\uff1a<\/p>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-28556\" title=\"Ollama \u5728 LangChain \u4e2d\u7684\u4f7f\u7528 - Python \u96c6\u6210-2\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/6f3c088dabe9da1.png\" alt=\"Ollama \u5728 LangChain \u4e2d\u7684\u4f7f\u7528 - Python \u96c6\u6210-2\" width=\"1186\" height=\"265\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/6f3c088dabe9da1.png 1186w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/6f3c088dabe9da1-768x172.png 768w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/6f3c088dabe9da1-18x4.png 18w\" sizes=\"auto, (max-width: 1186px) 100vw, 1186px\" \/><\/p>\n<p>\u6a21\u578b\u5b58\u50a8\u4f4d\u7f6e\uff1a<\/p>\n<ul>\n<li>Mac:\u00a0<code>~\/.ollama\/models\/<\/code><\/li>\n<li>Linux\uff08\u6216 WSL\uff09:\u00a0<code>\/usr\/share\/ollama\/.ollama\/models<\/code><\/li>\n<li>Windows:\u00a0<code>C:\\Users\\Administrator\\.ollama\\models<\/code><\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>3. \u57fa\u672c\u4f7f\u7528\u793a\u4f8b<\/h2>\n<h3>\u4f7f\u7528 ChatPromptTemplate \u8fdb\u884c\u5bf9\u8bdd<\/h3>\n<p>ChatPromptTemplate \u5141\u8bb8\u6211\u4eec\u521b\u5efa\u4e00\u4e2a\u53ef\u91cd\u7528\u7684\u6a21\u677f\uff0c\u5176\u4e2d\u5305\u542b\u4e00\u4e2a\u6216\u591a\u4e2a\u53c2\u6570\u3002\u8fd9\u4e9b\u53c2\u6570\u53ef\u4ee5\u5728\u8fd0\u884c\u65f6\u52a8\u6001\u66ff\u6362\uff0c\u4ee5\u751f\u6210\u4e0d\u540c\u7684\u63d0\u793a\u3002<\/p>\n<pre><code>template = \"\"\"\r\n\u4f60\u662f\u4e00\u4e2a\u4e50\u4e8e\u52a9\u4eba\u7684AI\uff0c\u64c5\u957f\u4e8e\u89e3\u51b3\u56de\u7b54\u5404\u79cd\u95ee\u9898\u3002\r\n\u95ee\u9898\uff1a{question}\r\n\"\"\"\r\nprompt = ChatPromptTemplate.from_template(template)\r\nchain = prompt | model\r\nchain.invoke({\"question\": \"\u4f60\u6bd4GPT4\u5389\u5bb3\u5417\uff1f\"})\r\n<\/code><\/pre>\n<p>\u5728\u521b\u5efa\u94fe\u90e8\u5206\uff0c\u4f7f\u7528\u7ba1\u9053\u64cd\u4f5c\u7b26\u00a0<code>|<\/code>\uff0c\u5b83\u5c06 prompt \u548c model \u8fde\u63a5\u8d77\u6765\uff0c\u5f62\u6210\u4e00\u4e2a\u5904\u7406\u6d41\u7a0b\u3002\u8fd9\u79cd\u94fe\u5f0f\u64cd\u4f5c\u4f7f\u5f97\u6211\u4eec\u53ef\u4ee5\u8f7b\u677e\u5730\u7ec4\u5408\u548c\u91cd\u7528\u4e0d\u540c\u7684\u7ec4\u4ef6\u3002<\/p>\n<p><code>invoke<\/code>\u00a0\u65b9\u6cd5\u89e6\u53d1\u6574\u4e2a\u5904\u7406\u94fe\uff0c\u5c06\u6211\u4eec\u7684\u95ee\u9898\u4f20\u5165\u6a21\u677f\uff0c\u7136\u540e\u5c06\u683c\u5f0f\u5316\u540e\u7684\u63d0\u793a\u53d1\u9001\u7ed9\u6a21\u578b\u8fdb\u884c\u5904\u7406\u3002<\/p>\n<h3>\u6d41\u5f0f\u8f93\u51fa<\/h3>\n<p>\u6d41\u5f0f\u8f93\u51fa\u662f\u4e00\u79cd\u5728\u751f\u6210\u957f\u6587\u672c\u65f6\u9010\u6b65\u8fd4\u56de\u7ed3\u679c\u7684\u6280\u672f\u3002\u8fd9\u79cd\u65b9\u6cd5\u6709\u51e0\u4e2a\u91cd\u8981\u7684\u4f18\u52bf\uff1a<\/p>\n<ol>\n<li>\u63d0\u9ad8\u7528\u6237\u4f53\u9a8c\uff1a\u7528\u6237\u53ef\u4ee5\u7acb\u5373\u770b\u5230\u90e8\u5206\u7ed3\u679c\uff0c\u800c\u4e0d\u662f\u7b49\u5f85\u6574\u4e2a\u54cd\u5e94\u5b8c\u6210\u3002<\/li>\n<li>\u51cf\u5c11\u7b49\u5f85\u65f6\u95f4\uff1a\u5bf9\u4e8e\u957f\u56de\u7b54\uff0c\u7528\u6237\u53ef\u4ee5\u5728\u5b8c\u6574\u56de\u7b54\u751f\u6210\u4e4b\u524d\u5c31\u5f00\u59cb\u9605\u8bfb\u3002<\/li>\n<li>\u5b9e\u65f6\u4ea4\u4e92\uff1a\u5141\u8bb8\u5728\u751f\u6210\u8fc7\u7a0b\u4e2d\u8fdb\u884c\u5e72\u9884\u6216\u7ec8\u6b62\u3002<\/li>\n<\/ol>\n<p>\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\uff0c\u7279\u522b\u662f\u5728\u804a\u5929\u673a\u5668\u4eba\u6216\u5b9e\u65f6\u5bf9\u8bdd\u7cfb\u7edf\u4e2d\uff0c\u6d41\u5f0f\u8f93\u51fa\u51e0\u4e4e\u662f\u5fc5\u4e0d\u53ef\u5c11\u7684\u3002<\/p>\n<pre><code>from langchain_ollama import <a href=\"https:\/\/www.kdjingpai.com\/en\/chatollamajiyun\/\">ChatOllama<\/a>\r\nmodel = ChatOllama(model=\"llama3.1\", temperature=0.7)\r\nmessages = [\r\n(\"human\", \"\u4f60\u597d\u5440\"),\r\n]\r\nfor chunk in model.stream(messages):\r\nprint(chunk.content, end='', flush=True)\r\n<\/code><\/pre>\n<p><code>model.stream()<\/code>\u00a0\u65b9\u6cd5\u662f\u5bf9 Ollama API \u7684\u6d41\u5f0f\u8f93\u51fa\u63a5\u53e3\u7684\u5c01\u88c5\uff0c\u5b83\u8fd4\u56de\u4e00\u4e2a\u751f\u6210\u5668\uff08generator\uff09\u5bf9\u8c61\u3002\u5f53\u8c03\u7528\u00a0<code>model.stream(messages)<\/code>\u00a0\u65f6\uff0c\u4f1a\u5b8c\u6210\u4ee5\u4e0b\u64cd\u4f5c\uff1a<\/p>\n<ul>\n<li>\u5411 Ollama API \u53d1\u9001\u8bf7\u6c42\uff0c\u5f00\u59cb\u751f\u6210\u54cd\u5e94\u3002<\/li>\n<li>API \u5f00\u59cb\u751f\u6210\u6587\u672c\uff0c\u4f46\u4e0d\u662f\u7b49\u5230\u5168\u90e8\u751f\u6210\u5b8c\u6bd5\u624d\u8fd4\u56de\uff0c\u800c\u662f\u4e00\u5c0f\u5757\u4e00\u5c0f\u5757\u5730\u8fd4\u56de\u3002<\/li>\n<li>\u6bcf\u6536\u5230\u4e00\u5c0f\u5757\u6587\u672c\uff0c<code>stream()<\/code>\u00a0\u65b9\u6cd5\u5c31\u4f1a yield \u8fd9\u4e2a\u6587\u672c\u5757\u3002<\/li>\n<li><code>flush=True<\/code>\u00a0\u786e\u4fdd\u6bcf\u4e2a\u7247\u6bb5\u7acb\u5373\u663e\u793a\uff0c\u800c\u4e0d\u662f\u7b49\u5f85\u7f13\u51b2\u533a\u586b\u6ee1\u3002<\/li>\n<\/ul>\n<h3>\u5de5\u5177\u8c03\u7528<\/h3>\n<p>\u5de5\u5177\u8c03\u7528\u662f AI \u6a21\u578b\u4e0e\u5916\u90e8\u51fd\u6570\u6216 API \u4ea4\u4e92\u7684\u80fd\u529b\u3002\u8fd9\u4f7f\u5f97\u6a21\u578b\u53ef\u4ee5\u6267\u884c\u590d\u6742\u7684\u4efb\u52a1\uff0c\u5982\u6570\u5b66\u8ba1\u7b97\u3001\u6570\u636e\u67e5\u8be2\u6216\u5916\u90e8\u670d\u52a1\u8c03\u7528\u3002<\/p>\n<pre><code>def simple_calculator(operation: str, x: float, y: float) -&gt; float:\r\n'''\u5b9e\u9645\u7684\u4ee3\u7801\u5904\u7406\u903b\u8f91'''\r\nllm = ChatOllama(\r\nmodel=\"llama3.1\",\r\ntemperature=0,\r\n).bind_tools([simple_calculator])\r\nresult = llm.invoke(\"\u4f60\u77e5\u9053\u4e00\u5343\u4e07\u4e58\u4e8c\u662f\u591a\u5c11\u5417\uff1f\")\r\n<\/code><\/pre>\n<p><code>bind_tools<\/code>\u00a0\u65b9\u6cd5\u5141\u8bb8\u6211\u4eec\u5c06\u81ea\u5b9a\u4e49\u51fd\u6570\u6ce8\u518c\u5230\u6a21\u578b\u4e2d\u3002\u8fd9\u6837\uff0c\u5f53\u6a21\u578b\u9047\u5230\u9700\u8981\u8ba1\u7b97\u7684\u95ee\u9898\u65f6\uff0c\u5b83\u53ef\u4ee5\u8c03\u7528\u8fd9\u4e2a\u51fd\u6570\u6765\u83b7\u5f97\u51c6\u786e\u7684\u7ed3\u679c\uff0c\u800c\u4e0d\u662f\u4f9d\u8d56\u4e8e\u5176\u9884\u8bad\u7ec3\u77e5\u8bc6\u3002<\/p>\n<p>\u8fd9\u79cd\u80fd\u529b\u5728\u6784\u5efa\u590d\u6742\u7684 AI \u5e94\u7528\u65f6\u975e\u5e38\u6709\u7528\uff0c\u4f8b\u5982\uff1a<\/p>\n<ul>\n<li>\u521b\u5efa\u53ef\u4ee5\u8bbf\u95ee\u5b9e\u65f6\u6570\u636e\u7684\u804a\u5929\u673a\u5668\u4eba<\/li>\n<li>\u6784\u5efa\u80fd\u6267\u884c\u7279\u5b9a\u4efb\u52a1\uff08\u5982\u9884\u8ba2\u3001\u67e5\u8be2\u7b49\uff09\u7684\u667a\u80fd\u52a9\u624b<\/li>\n<li>\u5f00\u53d1\u80fd\u8fdb\u884c\u7cbe\u786e\u8ba1\u7b97\u6216\u590d\u6742\u64cd\u4f5c\u7684 AI \u7cfb\u7edf<\/li>\n<\/ul>\n<h3>\u591a\u6a21\u6001\u6a21\u578b<\/h3>\n<p>Ollama \u652f\u6301\u591a\u6a21\u6001\u6a21\u578b\uff0c\u5982 bakllava \u548c llava\u3002\u591a\u6a21\u6001\u6a21\u578b\u662f\u80fd\u591f\u5904\u7406\u591a\u79cd\u7c7b\u578b\u8f93\u5165\uff08\u5982\u6587\u672c\u3001\u56fe\u50cf\u3001\u97f3\u9891\u7b49\uff09\u7684 AI \u6a21\u578b\u3002\u8fd9\u4e9b\u6a21\u578b\u5728\u7406\u89e3\u548c\u751f\u6210\u8de8\u6a21\u6001\u5185\u5bb9\u65b9\u9762\u8868\u73b0\u51fa\u8272\uff0c\u4f7f\u5f97\u66f4\u590d\u6742\u548c\u81ea\u7136\u7684\u4eba\u673a\u4ea4\u4e92\u6210\u4e3a\u53ef\u80fd\u3002<\/p>\n<p>\u9996\u5148\uff0c\u9700\u8981\u4e0b\u8f7d\u591a\u6a21\u6001\u6a21\u578b\u3002\u5728\u547d\u4ee4\u884c\u6267\u884c\uff1a<\/p>\n<pre><code>ollama pull llava\r\n<\/code><\/pre>\n<p><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter size-full wp-image-28555\" title=\"Ollama \u5728 LangChain \u4e2d\u7684\u4f7f\u7528 - Python \u96c6\u6210-3\" src=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/3795ae06a0c1294.png\" alt=\"Ollama \u5728 LangChain \u4e2d\u7684\u4f7f\u7528 - Python \u96c6\u6210-3\" width=\"1164\" height=\"260\" srcset=\"https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/3795ae06a0c1294.png 1164w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/3795ae06a0c1294-768x172.png 768w, https:\/\/www.kdjingpai.com\/wp-content\/uploads\/2025\/03\/3795ae06a0c1294-18x4.png 18w\" sizes=\"auto, (max-width: 1164px) 100vw, 1164px\" \/><\/p>\n<p>\u7136\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528\u4ee5\u4e0b\u4ee3\u7801\u6765\u5904\u7406\u56fe\u50cf\u548c\u6587\u672c\u8f93\u5165\uff1a<\/p>\n<pre><code>from langchain_ollama import ChatOllama\r\nfrom langchain_core.messages import HumanMessage\r\nfrom langchain_core.output_parsers import StrOutputParser\r\nllm = ChatOllama(model=\"llava\", temperature=0)\r\ndef prompt_func(data):\r\n'''\u6784\u9020\u591a\u6a21\u6001\u8f93\u5165'''\r\nchain = prompt_func | llm | StrOutputParser()\r\nquery_chain = chain.invoke(\r\n{\"text\": \"\u8fd9\u4e2a\u56fe\u7247\u91cc\u662f\u4ec0\u4e48\u52a8\u7269\u554a?\", \"image\": image_b64}\r\n)\r\n<\/code><\/pre>\n<p>\u8fd9\u91cc\u7684\u5173\u952e\u70b9\u662f\uff1a<\/p>\n<ol>\n<li>\u56fe\u50cf\u9884\u5904\u7406\uff1a\u6211\u4eec\u9700\u8981\u5c06\u56fe\u50cf\u8f6c\u6362\u4e3a base64 \u7f16\u7801\u7684\u5b57\u7b26\u4e32\u3002<\/li>\n<li>\u63d0\u793a\u51fd\u6570\uff1a<code>prompt_func<\/code>\u00a0\u521b\u5efa\u4e86\u4e00\u4e2a\u5305\u542b\u6587\u672c\u548c\u56fe\u50cf\u7684\u591a\u6a21\u6001\u8f93\u5165\u3002<\/li>\n<li>\u94fe\u5f0f\u5904\u7406\uff1a\u6211\u4eec\u4f7f\u7528\u00a0<code>|<\/code>\u00a0\u64cd\u4f5c\u7b26\u5c06\u63d0\u793a\u51fd\u6570\u3001\u6a21\u578b\u548c\u8f93\u51fa\u89e3\u6790\u5668\u8fde\u63a5\u8d77\u6765\u3002<\/li>\n<\/ol>\n<p>\u591a\u6a21\u6001\u6a21\u578b\u5728\u5f88\u591a\u573a\u666f\u4e0b\u90fd\u5f88\u6709\u7528\uff0c\u6bd4\u5982\uff1a<\/p>\n<ul>\n<li>\u56fe\u50cf\u63cf\u8ff0\u751f\u6210<\/li>\n<li>\u89c6\u89c9\u95ee\u7b54\u7cfb\u7edf<\/li>\n<li>\u57fa\u4e8e\u56fe\u50cf\u7684\u5185\u5bb9\u5206\u6790\u548c\u63a8\u8350<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>4. \u8fdb\u9636\u7528\u6cd5<\/h2>\n<h3>\u4f7f\u7528 ConversationChain \u8fdb\u884c\u5bf9\u8bdd<\/h3>\n<p><code>ConversationChain<\/code>\u00a0\u662f LangChain \u63d0\u4f9b\u7684\u4e00\u4e2a\u5f3a\u5927\u5de5\u5177\uff0c\u7528\u4e8e\u7ba1\u7406\u591a\u8f6e\u5bf9\u8bdd\u3002\u5b83\u7ed3\u5408\u4e86\u8bed\u8a00\u6a21\u578b\u3001\u63d0\u793a\u6a21\u677f\u548c\u5185\u5b58\u7ec4\u4ef6\uff0c\u4f7f\u5f97\u521b\u5efa\u5177\u6709\u4e0a\u4e0b\u6587\u611f\u77e5\u80fd\u529b\u7684\u5bf9\u8bdd\u7cfb\u7edf\u53d8\u5f97\u7b80\u5355\u3002<\/p>\n<pre><code>memory = ConversationBufferMemory()\r\nconversation = ConversationChain(\r\nllm=model,\r\nmemory=memory,\r\nverbose=True\r\n)\r\n# \u8fdb\u884c\u5bf9\u8bdd\r\nresponse = conversation.predict(input=\"\u4f60\u597d\uff0c\u6211\u60f3\u4e86\u89e3\u4e00\u4e0b\u4eba\u5de5\u667a\u80fd\u3002\")\r\nprint(\"AI:\", response)\r\nresponse = conversation.predict(input=\"\u80fd\u7ed9\u6211\u4e3e\u4e2aAI\u5728\u65e5\u5e38\u751f\u6d3b\u4e2d\u7684\u5e94\u7528\u4f8b\u5b50\u5417\uff1f\")\r\nprint(\"AI:\", response)\r\nresponse = conversation.predict(input=\"\u8fd9\u542c\u8d77\u6765\u5f88\u6709\u8da3\u3002AI\u5728\u533b\u7597\u9886\u57df\u6709\u4ec0\u4e48\u5e94\u7528\uff1f\")\r\nprint(\"AI:\", response)\r\n<\/code><\/pre>\n<p>\u8fd9\u91cc\u7684\u5173\u952e\u7ec4\u4ef6\u662f\uff1a<\/p>\n<ol>\n<li><code>ConversationBufferMemory<\/code>\uff1a\u8fd9\u662f\u4e00\u4e2a\u7b80\u5355\u7684\u5185\u5b58\u7ec4\u4ef6\uff0c\u5b83\u5b58\u50a8\u6240\u6709\u5148\u524d\u7684\u5bf9\u8bdd\u5386\u53f2\u3002<\/li>\n<li><code>ConversationChain<\/code>\uff1a\u5b83\u5c06\u8bed\u8a00\u6a21\u578b\u3001\u5185\u5b58\u548c\u4e00\u4e2a\u9ed8\u8ba4\u7684\u5bf9\u8bdd\u63d0\u793a\u6a21\u677f\u7ec4\u5408\u5728\u4e00\u8d77\u3002<\/li>\n<\/ol>\n<p>\u7ef4\u62a4\u5bf9\u8bdd\u5386\u53f2\u5f88\u91cd\u8981\uff0c\u56e0\u4e3a\u5b83\u5141\u8bb8\u6a21\u578b\uff1a<\/p>\n<ul>\n<li>\u7406\u89e3\u4e0a\u4e0b\u6587\u548c\u4e4b\u524d\u63d0\u5230\u7684\u4fe1\u606f<\/li>\n<li>\u751f\u6210\u66f4\u8fde\u8d2f\u548c\u76f8\u5173\u7684\u56de\u590d<\/li>\n<li>\u5904\u7406\u590d\u6742\u7684\u591a\u8f6e\u5bf9\u8bdd\u573a\u666f<\/li>\n<\/ul>\n<p>\u5728\u5b9e\u9645\u5e94\u7528\u4e2d\uff0c\u4f60\u53ef\u80fd\u9700\u8981\u8003\u8651\u4f7f\u7528\u66f4\u9ad8\u7ea7\u7684\u5185\u5b58\u7ec4\u4ef6\uff0c\u5982\u00a0<code>ConversationSummaryMemory<\/code>\uff0c\u4ee5\u5904\u7406\u957f\u5bf9\u8bdd\u5e76\u907f\u514d\u8d85\u51fa\u6a21\u578b\u7684\u4e0a\u4e0b\u6587\u957f\u5ea6\u9650\u5236\u3002<\/p>\n<h3>\u81ea\u5b9a\u4e49\u63d0\u793a\u6a21\u677f<\/h3>\n<p>\u8bbe\u8ba1\u597d\u7684\u63d0\u793a\u6a21\u677f\u662f\u521b\u5efa\u9ad8\u6548 AI \u5e94\u7528\u7684\u5173\u952e\u3002\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u521b\u5efa\u4e86\u4e00\u4e2a\u7528\u4e8e\u751f\u6210\u4ea7\u54c1\u63cf\u8ff0\u7684\u590d\u6742\u63d0\u793a\uff1a<\/p>\n<pre><code>system_message = SystemMessage(content=\"\"\"\r\n\u4f60\u662f\u4e00\u4f4d\u7ecf\u9a8c\u4e30\u5bcc\u7684\u7535\u5546\u6587\u6848\u64b0\u5199\u4e13\u5bb6\u3002\u4f60\u7684\u4efb\u52a1\u662f\u6839\u636e\u7ed9\u5b9a\u7684\u4ea7\u54c1\u4fe1\u606f\u521b\u4f5c\u5438\u5f15\u4eba\u7684\u5546\u54c1\u63cf\u8ff0\u3002\r\n\u8bf7\u786e\u4fdd\u4f60\u7684\u63cf\u8ff0\u7b80\u6d01\u3001\u6709\u529b\uff0c\u5e76\u4e14\u7a81\u51fa\u4ea7\u54c1\u7684\u6838\u5fc3\u4f18\u52bf\u3002\r\n\"\"\")\r\nhuman_message_template = \"\"\"\r\n\u8bf7\u4e3a\u4ee5\u4e0b\u4ea7\u54c1\u521b\u4f5c\u4e00\u6bb5\u5438\u5f15\u4eba\u7684\u5546\u54c1\u63cf\u8ff0\uff1a\r\n\u4ea7\u54c1\u7c7b\u578b: {product_type}\r\n\u6838\u5fc3\u7279\u6027: {key_feature}\r\n\u76ee\u6807\u53d7\u4f17: {target_audience}\r\n\u4ef7\u683c\u533a\u95f4: {price_range}\r\n\u54c1\u724c\u5b9a\u4f4d: {brand_positioning}\r\n\u8bf7\u63d0\u4f9b\u4ee5\u4e0b\u4e09\u79cd\u4e0d\u540c\u98ce\u683c\u7684\u63cf\u8ff0\uff0c\u6bcf\u79cd\u5927\u7ea650\u5b57\uff1a\r\n1. \u7406\u6027\u5206\u6790\u578b\r\n2. \u60c5\u611f\u8bc9\u6c42\u578b\r\n3. \u6545\u4e8b\u5316\u8425\u9500\u578b\r\n\"\"\"\r\n# \u793a\u4f8b\u4f7f\u7528\r\nproduct_info = {\r\n\"product_type\": \"\u667a\u80fd\u624b\u8868\",\r\n\"key_feature\": \"\u5fc3\u7387\u76d1\u6d4b\u548c\u7761\u7720\u5206\u6790\",\r\n\"target_audience\": \"\u6ce8\u91cd\u5065\u5eb7\u7684\u5e74\u8f7b\u4e13\u4e1a\u4eba\u58eb\",\r\n\"price_range\": \"\u4e2d\u9ad8\u7aef\",\r\n\"brand_positioning\": \"\u79d1\u6280\u4e0e\u5065\u5eb7\u7684\u5b8c\u7f8e\u7ed3\u5408\"\r\n}\r\n<\/code><\/pre>\n<p>\u8fd9\u4e2a\u7ed3\u6784\u6709\u51e0\u4e2a\u91cd\u8981\u7684\u8bbe\u8ba1\u8003\u8651\uff1a<\/p>\n<ol>\n<li>system_prompt\uff1a\u5b9a\u4e49\u4e86 AI \u7684\u89d2\u8272\u548c\u603b\u4f53\u4efb\u52a1\uff0c\u8bbe\u7f6e\u4e86\u6574\u4e2a\u5bf9\u8bdd\u7684\u57fa\u8c03\u3002<\/li>\n<li>human_message_template\uff1a\u63d0\u4f9b\u4e86\u5177\u4f53\u7684\u6307\u4ee4\u548c\u6240\u9700\u4fe1\u606f\u7684\u7ed3\u6784\u3002<\/li>\n<li>\u591a\u53c2\u6570\u8bbe\u8ba1\uff1a\u5141\u8bb8\u7075\u6d3b\u5730\u9002\u5e94\u4e0d\u540c\u7684\u4ea7\u54c1\u548c\u9700\u6c42\u3002<\/li>\n<li>\u591a\u6837\u5316\u8f93\u51fa\u8981\u6c42\uff1a\u901a\u8fc7\u8981\u6c42\u4e0d\u540c\u98ce\u683c\u7684\u63cf\u8ff0\uff0c\u9f13\u52b1\u6a21\u578b\u5c55\u793a\u5176\u591a\u6837\u6027\u3002<\/li>\n<\/ol>\n<p>\u8bbe\u8ba1\u6709\u6548\u7684\u63d0\u793a\u6a21\u677f\u65f6\uff0c\u8003\u8651\u4ee5\u4e0b\u51e0\u70b9\uff1a<\/p>\n<ul>\n<li>\u660e\u786e\u5b9a\u4e49 AI \u7684\u89d2\u8272\u548c\u4efb\u52a1<\/li>\n<li>\u63d0\u4f9b\u6e05\u6670\u3001\u7ed3\u6784\u5316\u7684\u8f93\u5165\u683c\u5f0f<\/li>\n<li>\u5305\u542b\u5177\u4f53\u7684\u8f93\u51fa\u8981\u6c42\u548c\u683c\u5f0f\u6307\u5bfc<\/li>\n<li>\u8003\u8651\u5982\u4f55\u6700\u5927\u5316\u6a21\u578b\u7684\u80fd\u529b\u548c\u521b\u9020\u529b<\/li>\n<\/ul>\n<h3>\u6784\u5efa\u4e00\u4e2a\u7b80\u5355\u7684 RAG \u95ee\u7b54\u7cfb\u7edf<\/h3>\n<p>RAG\uff08Retrieval-Augmented Generation\uff09\u662f\u4e00\u79cd\u7ed3\u5408\u4e86\u68c0\u7d22\u548c\u751f\u6210\u7684 AI \u6280\u672f\uff0c\u5b83\u901a\u8fc7\u68c0\u7d22\u76f8\u5173\u4fe1\u606f\u6765\u589e\u5f3a\u8bed\u8a00\u6a21\u578b\u7684\u56de\u7b54\u80fd\u529b\u3002RAG \u7cfb\u7edf\u7684\u5de5\u4f5c\u6d41\u7a0b\u901a\u5e38\u5305\u62ec\u4ee5\u4e0b\u6b65\u9aa4\uff1a<\/p>\n<ol>\n<li>\u5c06\u77e5\u8bc6\u5e93\u6587\u6863\u5206\u5272\u6210\u5c0f\u5757\u5e76\u521b\u5efa\u5411\u91cf\u7d22\u5f15<\/li>\n<li>\u5bf9\u7528\u6237\u7684\u95ee\u9898\u8fdb\u884c\u5411\u91cf\u5316\uff0c\u5728\u7d22\u5f15\u4e2d\u68c0\u7d22\u76f8\u5173\u6587\u6863<\/li>\n<li>\u5c06\u68c0\u7d22\u5230\u7684\u76f8\u5173\u6587\u6863\u548c\u539f\u59cb\u95ee\u9898\u4e00\u8d77\u4f5c\u4e3a\u4e0a\u4e0b\u6587\u63d0\u4f9b\u7ed9\u8bed\u8a00\u6a21\u578b<\/li>\n<li>\u8bed\u8a00\u6a21\u578b\u6839\u636e\u68c0\u7d22\u5230\u7684\u4fe1\u606f\u751f\u6210\u56de\u7b54<\/li>\n<\/ol>\n<p>RAG \u7684\u4f18\u52bf\u5728\u4e8e\u5b83\u53ef\u4ee5\u5e2e\u52a9\u8bed\u8a00\u6a21\u578b\u8bbf\u95ee\u6700\u65b0\u548c\u4e13\u4e1a\u7684\u4fe1\u606f\uff0c\u51cf\u5c11\u5e7b\u89c9\uff0c\u5e76\u63d0\u9ad8\u56de\u7b54\u7684\u51c6\u786e\u6027\u548c\u76f8\u5173\u6027\u3002<\/p>\n<p>LangChain \u63d0\u4f9b\u4e86\u591a\u79cd\u7ec4\u4ef6\uff0c\u53ef\u4ee5\u4e0e Ollama \u6a21\u578b\u65e0\u7f1d\u96c6\u6210\u3002\u8fd9\u91cc\u6211\u4eec\u5c06\u5c55\u793a\u5982\u4f55\u5c06 Ollama \u6a21\u578b\u4e0e\u5411\u91cf\u5b58\u50a8\u548c\u68c0\u7d22\u5668\u7ed3\u5408\u4f7f\u7528\uff0c\u521b\u5efa\u4e00\u4e2a\u7b80\u5355\u7684 RAG \u95ee\u7b54\u7cfb\u7edf\u3002<\/p>\n<p>\u9996\u5148\u9700\u8981\u786e\u4fdd\u4e0b\u8f7d embedding \u6a21\u578b\uff0c\u53ef\u4ee5\u5728\u547d\u4ee4\u884c\u6267\u884c\u4ee5\u4e0b\u547d\u4ee4\uff1a<\/p>\n<pre><code>ollama pull nomic-embed-text\r\n<\/code><\/pre>\n<p>\u7136\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u6784\u5efa RAG \u7cfb\u7edf\uff1a<\/p>\n<pre><code># \u521d\u59cb\u5316 Ollama \u6a21\u578b\u548c\u5d4c\u5165\r\nllm = ChatOllama(model=\"llama3.1\")\r\nembeddings = OllamaEmbeddings(model=\"nomic-embed-text\")\r\n# \u51c6\u5907\u6587\u6863\r\ntext = \"\"\"\r\nDatawhale \u662f\u4e00\u4e2a\u4e13\u6ce8\u4e8e\u6570\u636e\u79d1\u5b66\u4e0e AI \u9886\u57df\u7684\u5f00\u6e90\u7ec4\u7ec7\uff0c\u6c47\u96c6\u4e86\u4f17\u591a\u9886\u57df\u9662\u6821\u548c\u77e5\u540d\u4f01\u4e1a\u7684\u4f18\u79c0\u5b66\u4e60\u8005\uff0c\u805a\u5408\u4e86\u4e00\u7fa4\u6709\u5f00\u6e90\u7cbe\u795e\u548c\u63a2\u7d22\u7cbe\u795e\u7684\u56e2\u961f\u6210\u5458\u3002\r\nDatawhale \u4ee5\" for the learner\uff0c\u548c\u5b66\u4e60\u8005\u4e00\u8d77\u6210\u957f\"\u4e3a\u613f\u666f\uff0c\u9f13\u52b1\u771f\u5b9e\u5730\u5c55\u73b0\u81ea\u6211\u3001\u5f00\u653e\u5305\u5bb9\u3001\u4e92\u4fe1\u4e92\u52a9\u3001\u6562\u4e8e\u8bd5\u9519\u548c\u52c7\u4e8e\u62c5\u5f53\u3002\r\n\u540c\u65f6 Datawhale \u7528\u5f00\u6e90\u7684\u7406\u5ff5\u53bb\u63a2\u7d22\u5f00\u6e90\u5185\u5bb9\u3001\u5f00\u6e90\u5b66\u4e60\u548c\u5f00\u6e90\u65b9\u6848\uff0c\u8d4b\u80fd\u4eba\u624d\u57f9\u517b\uff0c\u52a9\u529b\u4eba\u624d\u6210\u957f\uff0c\u5efa\u7acb\u8d77\u4eba\u4e0e\u4eba\uff0c\u4eba\u4e0e\u77e5\u8bc6\uff0c\u4eba\u4e0e\u4f01\u4e1a\u548c\u4eba\u4e0e\u672a\u6765\u7684\u8054\u7ed3\u3002\r\n\u5982\u679c\u4f60\u60f3\u5728Datawhale\u5f00\u6e90\u793e\u533a\u53d1\u8d77\u4e00\u4e2a\u5f00\u6e90\u9879\u76ee\uff0c\u8bf7\u8be6\u7ec6\u9605\u8bfbDatawhale\u5f00\u6e90\u9879\u76ee\u6307\u5357[https:\/\/github.com\/datawhalechina\/DOPMC\/blob\/main\/GUIDE.md]\r\n\"\"\"\r\n# \u5206\u5272\u6587\u672c\r\ntext_splitter = RecursiveCharacterTextSplitter(chunk_size=100, chunk_overlap=20)\r\nchunks = text_splitter.split_text(text)\r\n# \u521b\u5efa\u5411\u91cf\u5b58\u50a8\r\nvectorstore = FAISS.from_texts(chunks, embeddings)\r\nretriever = vectorstore.as_retriever()\r\n# \u521b\u5efa\u63d0\u793a\u6a21\u677f\r\ntemplate = \"\"\"\u53ea\u80fd\u4f7f\u7528\u4e0b\u5217\u5185\u5bb9\u56de\u7b54\u95ee\u9898:\r\n{context}\r\nQuestion: {question}\r\n\"\"\"\r\nprompt = ChatPromptTemplate.from_template(template)\r\n# \u521b\u5efa\u68c0\u7d22-\u95ee\u7b54\u94fe\r\nchain = (\r\n{\"context\": retriever, \"question\": RunnablePassthrough()}\r\n| prompt\r\n| llm\r\n)\r\n# \u4f7f\u7528\u94fe\u56de\u7b54\u95ee\u9898\r\nquestion = \"\u6211\u60f3\u4e3adatawhale\u8d21\u732e\u8be5\u600e\u4e48\u505a\uff1f\"\r\nresponse = chain.invoke(question)\r\n<\/code><\/pre>\n<p>\u8fd9\u4e2a RAG \u7cfb\u7edf\u7684\u5de5\u4f5c\u539f\u7406\u5982\u4e0b\uff1a<\/p>\n<ol>\n<li>\u6587\u672c\u5206\u5272\uff1a\u4f7f\u7528\u00a0<code>RecursiveCharacterTextSplitter<\/code>\u00a0\u5c06\u957f\u6587\u672c\u5206\u5272\u6210\u5c0f\u5757\u3002<\/li>\n<li>\u5411\u91cf\u5316\u548c\u7d22\u5f15\uff1a\u4f7f\u7528\u00a0<code>OllamaEmbeddings<\/code>\u00a0\u5c06\u6587\u672c\u5757\u8f6c\u6362\u4e3a\u5411\u91cf\uff0c\u5e76\u7528 FAISS \u521b\u5efa\u5411\u91cf\u7d22\u5f15\u3002<\/li>\n<li>\u68c0\u7d22\uff1a\u5f53\u6536\u5230\u95ee\u9898\u65f6\uff0c\u7cfb\u7edf\u4f1a\u5c06\u95ee\u9898\u5411\u91cf\u5316\uff0c\u5e76\u5728 FAISS \u7d22\u5f15\u4e2d\u68c0\u7d22\u6700\u76f8\u5173\u7684\u6587\u672c\u5757\u3002<\/li>\n<li>\u751f\u6210\u56de\u7b54\uff1a\u5c06\u68c0\u7d22\u5230\u7684\u76f8\u5173\u6587\u672c\u5757\u4e0e\u539f\u59cb\u95ee\u9898\u4e00\u8d77\u63d0\u4f9b\u7ed9\u8bed\u8a00\u6a21\u578b\uff0c\u751f\u6210\u6700\u7ec8\u7b54\u6848\u3002<\/li>\n<\/ol>\n<p>RAG \u7cfb\u7edf\u5728\u8bb8\u591a\u771f\u5b9e\u573a\u666f\u4e2d\u975e\u5e38\u6709\u7528\uff0c\u4f8b\u5982\uff1a<\/p>\n<ul>\n<li>\u5ba2\u6237\u670d\u52a1\uff1a\u53ef\u4ee5\u57fa\u4e8e\u516c\u53f8\u7684\u77e5\u8bc6\u5e93\u5feb\u901f\u56de\u7b54\u5ba2\u6237\u8be2\u95ee\u3002<\/li>\n<li>\u7814\u7a76\u8f85\u52a9\uff1a\u5e2e\u52a9\u7814\u7a76\u4eba\u5458\u5feb\u901f\u627e\u5230\u76f8\u5173\u6587\u732e\u5e76\u603b\u7ed3\u5173\u952e\u4fe1\u606f\u3002<\/li>\n<li>\u4e2a\u4eba\u52a9\u624b\uff1a\u7ed3\u5408\u4e2a\u4eba\u7b14\u8bb0\u548c\u7f51\u7edc\u4fe1\u606f\uff0c\u63d0\u4f9b\u4e2a\u6027\u5316\u7684\u4fe1\u606f\u68c0\u7d22\u548c\u5efa\u8bae\u3002<\/li>\n<\/ul>\n<p>&nbsp;<\/p>\n<h2>\u7ed3\u8bba<\/h2>\n<p>\u901a\u8fc7\u8fd9\u4e9b\u793a\u4f8b\uff0c\u6211\u4eec\u5c55\u793a\u4e86\u5982\u4f55\u4f7f\u7528 Ollama \u548c LangChain \u6784\u5efa\u5404\u79cd AI \u5e94\u7528\uff0c\u4ece\u7b80\u5355\u7684\u5bf9\u8bdd\u7cfb\u7edf\u5230\u590d\u6742\u7684 RAG \u95ee\u7b54\u7cfb\u7edf\u3002\u8fd9\u4e9b\u5de5\u5177\u548c\u6280\u672f\u4e3a\u5f00\u53d1\u5f3a\u5927\u7684 AI \u5e94\u7528\u63d0\u4f9b\u4e86\u575a\u5b9e\u7684\u57fa\u7840\u3002<\/p>\n<p>Ollama \u548c LangChain \u7684\u7ed3\u5408\u4e3a\u5f00\u53d1\u8005\u63d0\u4f9b\u4e86\u6781\u5927\u7684\u7075\u6d3b\u6027\u548c\u53ef\u80fd\u6027\u3002\u4f60\u53ef\u4ee5\u6839\u636e\u5177\u4f53\u9700\u6c42\uff0c\u9009\u62e9\u5408\u9002\u7684\u6a21\u578b\u548c\u7ec4\u4ef6\uff0c\u6784\u5efa\u51fa\u9002\u5408\u4f60\u7684\u5e94\u7528\u573a\u666f\u7684 AI \u7cfb\u7edf\u3002<\/p>\n<p>\u968f\u7740\u6280\u672f\u7684\u4e0d\u65ad\u53d1\u5c55\uff0c\u6211\u4eec\u671f\u5f85\u770b\u5230\u66f4\u591a\u521b\u65b0\u7684\u5e94\u7528\u51fa\u73b0\u3002\u5e0c\u671b\u8fd9\u4e2a\u6307\u5357\u80fd\u591f\u5e2e\u52a9\u4f60\u5f00\u59cb\u4f60\u7684 AI \u5f00\u53d1\u4e4b\u65c5\uff0c\u5e76\u6fc0\u53d1\u4f60\u7684\u521b\u9020\u529b\uff0c\u53bb\u63a2\u7d22 AI \u6280\u672f\u7684\u65e0\u9650\u53ef\u80fd\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u7b80\u4ecb \u672c\u6587\u6863\u4ecb\u7ecd\u4e86\u5982\u4f55\u5728 Python \u73af\u5883\u4e2d\u4f7f\u7528 Ollama \u4e0e LangChain \u96c6\u6210\uff0c\u4ee5\u521b\u5efa\u5f3a\u5927\u7684 AI \u5e94\u7528\u3002Ollama \u662f\u4e00\u4e2a\u5f00\u6e90\u7684\u5927\u8bed\u8a00\u6a21\u578b\u90e8\u7f72\u5de5\u5177\uff0c\u800c LangChain \u5219\u662f\u4e00\u4e2a\u7528\u4e8e\u6784\u5efa\u57fa\u4e8e\u8bed\u8a00\u6a21\u578b\u7684\u5e94\u7528\u7684\u6846\u67b6\u3002\u901a\u8fc7\u7ed3&#8230;<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[182],"tags":[],"class_list":["post-28553","post","type-post","status-publish","format-standard","hentry","category-shicao"],"_links":{"self":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts\/28553","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/comments?post=28553"}],"version-history":[{"count":0,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/posts\/28553\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/media?parent=28553"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/categories?post=28553"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.kdjingpai.com\/en\/wp-json\/wp\/v2\/tags?post=28553"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}