{"id":3622,"date":"2024-05-07T20:33:03","date_gmt":"2024-05-07T12:33:03","guid":{"rendered":"https:\/\/www.aqwu.net\/wp\/?p=3622"},"modified":"2024-05-07T21:32:58","modified_gmt":"2024-05-07T13:32:58","slug":"rag-%e5%85%a5%e9%97%a8%e6%95%99%e7%a8%8bqwen","status":"publish","type":"post","link":"https:\/\/www.aqwu.net\/wp\/?p=3622","title":{"rendered":"RAG \u5165\u95e8\u6559\u7a0b(Qwen)"},"content":{"rendered":"\n<p>\u672c\u6559\u7a0b\u4f7f\u7528\u4e86<a href=\"https:\/\/huggingface.co\/Qwen\/Qwen1.5-0.5B-Chat\">Qwen1.5-0.5B-Chat<\/a> \u4f5c\u4e3a\u5d4c\u5165\u7684\u5c0f\u6a21\u578b<\/p>\n\n\n\n<p>\u4f7f\u7528 <a href=\"https:\/\/huggingface.co\/Qwen\/Qwen1.5-1.8B-Chat\">Qwen1.5-1.8B-Chat<\/a> \u4f5c\u4e3a\u63a8\u7406\u6a21\u578b<\/p>\n\n\n\n<p>\u5f53\u7136\uff0c\u4f60\u4e5f\u53ef\u4ee5\u4f7f\u7528\u540c\u4e00\u4e2a\u6a21\u578b\u6216\u662f\u5176\u4ed6\u6a21\u578b<\/p>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>1. \u5d4c\u5165\u6a21\u578b<\/strong><\/h2>\n\n\n\n<p>\u6784\u5efa\u6587\u672c\u5d4c\u5165\u7684\u6a21\u578b\u6709\u5f88\u591a\uff0c\u4e0d\u540c\u7684\u6a21\u578b\u9002\u7528\u4e8e\u4e0d\u540c\u7684\u4efb\u52a1\u548c\u9700\u6c42\u3002\u5728\u9009\u62e9\u6a21\u578b\u65f6\uff0c\u60a8\u5e94\u8be5\u8003\u8651\u60a8\u7684\u5177\u4f53\u5e94\u7528\u573a\u666f\uff0c\u6bd4\u5982\u5904\u7406\u8bed\u8a00\u7684\u79cd\u7c7b\u3001\u4efb\u52a1\u7684\u590d\u6742\u6027\u3001\u4ee5\u53ca\u60a8\u7684\u7cfb\u7edf\u8d44\u6e90\u3002\u4ee5\u4e0b\u662f\u4e00\u4e9b\u5e38\u7528\u4e8e\u6784\u5efa\u6587\u672c\u5d4c\u5165\u7684\u6a21\u578b\u7c7b\u522b\u548c\u5177\u4f53\u6a21\u578b\uff1a<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>1.1. \u9884\u8bad\u7ec3\u8bed\u8a00\u6a21\u578b\uff08Pre-trained Language Models\uff09<\/strong><\/h3>\n\n\n\n<p>\u8fd9\u4e9b\u6a21\u578b\u901a\u5e38\u57fa\u4e8e\u5927\u89c4\u6a21\u6587\u672c\u6570\u636e\u96c6\u9884\u8bad\u7ec3\uff0c\u80fd\u591f\u6355\u6349\u4e30\u5bcc\u7684\u8bed\u8a00\u7279\u5f81\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>BERT\uff08Bidirectional Encoder Representations from Transformers\uff09<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u9002\u7528\u4e8e\u5e7f\u6cdb\u7684\u81ea\u7136\u8bed\u8a00\u5904\u7406\u4efb\u52a1\uff0c\u80fd\u591f\u6355\u83b7\u4e0a\u4e0b\u6587\u4e2d\u7684\u53cc\u5411\u5173\u7cfb\u3002<\/li>\n\n\n\n<li>\u53d8\u79cd\u5305\u62ec RoBERTa\u3001DistilBERT\u3001ALBERT \u7b49\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>GPT\uff08Generative Pre-trained Transformer\uff09<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u57fa\u4e8e Transformer \u7684\u81ea\u56de\u5f52\u6a21\u578b\uff0c\u9002\u5408\u751f\u6210\u4efb\u52a1\u3002<\/li>\n\n\n\n<li>\u5404\u79cd\u7248\u672c\u5982 GPT-2, GPT-3 \u7b49\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>T5\uff08Text-To-Text Transfer Transformer\uff09<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u5c06\u6240\u6709\u6587\u672c\u4efb\u52a1\u8f6c\u6362\u4e3a\u6587\u672c\u5230\u6587\u672c\u7684\u95ee\u9898\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>XLNet<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u7ed3\u5408\u4e86\u81ea\u56de\u5f52\u548c\u81ea\u7f16\u7801\u7279\u70b9\uff0c\u4f18\u4e8e BERT \u7684\u6027\u80fd\u5728\u67d0\u4e9b\u4efb\u52a1\u4e0a\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>1.2. \u53e5\u5b50\u5d4c\u5165\u6a21\u578b\uff08Sentence Embedding Models\uff09<\/strong><\/h3>\n\n\n\n<p>\u4e13\u95e8\u8bbe\u8ba1\u7528\u6765\u76f4\u63a5\u8f93\u51fa\u53e5\u5b50\u6216\u6bb5\u843d\u7ea7\u522b\u7684\u5d4c\u5165\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Sentence-BERT\uff08SBERT\uff09<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u4e00\u4e2a BERT \u7684\u4fee\u6539\u7248\u672c\uff0c\u4f18\u5316\u4e86\u53e5\u5b50\u7ea7\u522b\u7684\u76f8\u4f3c\u6027\u6bd4\u8f83\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>Universal Sentence Encoder<\/strong>\n<ul class=\"wp-block-list\">\n<li>Google \u5f00\u53d1\uff0c\u652f\u6301\u591a\u79cd\u8bed\u8a00\uff0c\u9002\u5408\u5927\u8303\u56f4\u7684\u4efb\u52a1\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>1.3. \u4e13\u95e8\u5316\u5d4c\u5165\u6a21\u578b<\/strong><\/h3>\n\n\n\n<p>\u9488\u5bf9\u7279\u5b9a\u5e94\u7528\u6216\u9886\u57df\u4f18\u5316\u7684\u6a21\u578b\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>FastText<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u7531 Facebook \u5f00\u53d1\uff0c\u7279\u522b\u9002\u7528\u4e8e\u5305\u62ec\u7a00\u6709\u8bcd\u6c47\u5728\u5185\u7684\u6587\u672c\u5904\u7406\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>ELMo\uff08Embeddings from Language Models\uff09<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u4f7f\u7528\u53cc\u5411 LSTM \u7f51\u7edc\u7ed3\u6784\uff0c\u53ef\u4ee5\u751f\u6210\u4e0a\u4e0b\u6587\u76f8\u5173\u7684\u8bcd\u5d4c\u5165\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>1.4. \u9002\u5e94\u6027\u5d4c\u5165\u6a21\u578b<\/strong><\/h3>\n\n\n\n<p>\u80fd\u591f\u901a\u8fc7\u5fae\u8c03\u9002\u5e94\u7279\u5b9a\u4efb\u52a1\u6216\u9886\u57df\u7684\u6a21\u578b\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>AdaptBERT<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u901a\u8fc7\u5bf9\u7279\u5b9a\u9886\u57df\u6570\u636e\u7684\u5fae\u8c03\uff0c\u63d0\u9ad8\u6a21\u578b\u5728\u8be5\u9886\u57df\u7684\u6027\u80fd\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>1.5. \u591a\u8bed\u8a00\u548c\u8de8\u8bed\u8a00\u6a21\u578b<\/strong><\/h3>\n\n\n\n<p>\u652f\u6301\u591a\u79cd\u8bed\u8a00\uff0c\u9002\u5408\u8de8\u8bed\u8a00\u7684\u5e94\u7528\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>mBERT\uff08Multilingual BERT\uff09<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u652f\u6301\u591a\u79cd\u8bed\u8a00\u7684 BERT \u7248\u672c\u3002<\/li>\n<\/ul>\n<\/li>\n\n\n\n<li><strong>XLM<\/strong>\n<ul class=\"wp-block-list\">\n<li>\u9488\u5bf9\u8de8\u8bed\u8a00\u7406\u89e3\u4f18\u5316\u7684\u6a21\u578b\u3002<\/li>\n<\/ul>\n<\/li>\n<\/ul>\n\n\n\n<p>\u9009\u62e9\u6a21\u578b\u65f6\uff0c\u9664\u4e86\u8003\u8651\u6a21\u578b\u7684\u6027\u80fd\uff0c\u8fd8\u9700\u8981\u8003\u8651\u5b9e\u73b0\u7684\u590d\u6742\u5ea6\u3001\u8fd0\u884c\u65f6\u7684\u8d44\u6e90\u9700\u6c42\u3001\u4ee5\u53ca\u662f\u5426\u9700\u8981\u652f\u6301\u7279\u5b9a\u7684\u8bed\u8a00\u6216\u9886\u57df\u3002\u4e00\u822c\u6765\u8bf4\uff0c\u9884\u8bad\u7ec3\u6a21\u578b\u80fd\u63d0\u4f9b\u826f\u597d\u7684\u901a\u7528\u6027\u548c\u6027\u80fd\uff0c\u4f46\u5728\u7279\u5b9a\u60c5\u51b5\u4e0b\u53ef\u80fd\u9700\u8981\u8fdb\u4e00\u6b65\u7684\u8c03\u6574\u6216\u4f18\u5316\u3002<\/p>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>2. \u5b89\u88c5\u5fc5\u8981\u7684\u5e93<\/strong><\/h2>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:sh decode:true \">!python -m pip install --upgrade pip\n!pip install -U torch\n!pip install -U transformers\n!pip install -U sentence_transformers\n!pip install -U numpy\n!pip install -U faiss-cpu \n#!pip install faiss-gpu\n<\/pre><\/div>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>3. \u5bfc\u5165\u5e93<\/strong><\/h2>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:python decode:true \">import torch\nfrom transformers import AutoTokenizer, AutoModelForCausalLM\nfrom sentence_transformers import SentenceTransformer\nimport faiss\nimport numpy as np<\/pre><\/div>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>4. \u52a0\u8f7d\u6a21\u578b<\/strong><\/h2>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:python decode:true \" ># \u5c0f\u6a21\u578b\u7528\u4e8e\u521b\u5efa\u5d4c\u5165\nembedder = SentenceTransformer('Qwen\/Qwen1.5-0.5B-Chat')\n# \u652f\u6301\u591a\u8bed\u8a00\n# embedder = SentenceTransformer('bert-base-multilingual-cased')\n\n# \u5927\u6a21\u578b\u7528\u4e8e\u751f\u6210\ntokenizer = AutoTokenizer.from_pretrained('Qwen\/Qwen1.5-1.8B-Chat')\n\ndevice = \"cpu\" # the device to load the model onto\n\nmodel = AutoModelForCausalLM.from_pretrained(\n    \"Qwen\/Qwen1.5-1.8B-Chat\",\n    torch_dtype=\"auto\",\n    device_map=\"cpu\"\n)<\/pre><\/div>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>5. \u52a0\u8f7d\u6587\u4ef6<\/strong><\/h2>\n\n\n\n<p>\u8fd9\u4e2a\u4f8b\u5b50\u4f7f\u7528\u4e86\u4fdd\u7f57\u00b7\u683c\u96f7\u5384\u59c6\uff08Paul Graham\uff09\u7684\u6587\u7ae0\u201c<a href=\"http:\/\/paulgraham.com\/worked.html\">\u6211\u505a\u4e86\u4ec0\u4e48<\/a>\u201d\u7684\u6587\u672c\u3002<\/p>\n\n\n\n<p>\u83b7\u53d6\u5b83\u7684\u6700\u7b80\u5355\u65b9\u6cd5\u662f\u901a\u8fc7\u6b64\u94fe\u63a5\u4e0b\u8f7d\u5b83\u5e76\u5c06\u5176\u4fdd\u5b58\u5728&nbsp;<code>data<\/code> \u76ee\u5f55\u4e0b\uff0c\u6587\u4ef6\u540d\u4e3a\uff1apaul_graham_essay.txt<\/p>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:python decode:true \"># Open the file\nfile_path = \"data\/paul_graham_essay.txt\"\n\n# Initialize an empty list to store the embeddings\nembeddings_list = []\ndocuments = []\n\nwith open(file_path, \"r\", encoding=\"utf-8\") as file:\n    document = \"\"\n    for line in file:\n        # \u5f53\u7d2f\u79ef\u8d85\u8fc71000\u5b57\u7b26\u6216\u9047\u5230\u7a7a\u884c\u65f6\uff0c\u5904\u7406\u6587\u672c\n        if len(document) + len(line) &gt; 1000 or line.strip() == \"\":\n            document = document.strip()\n            if document:  # \u786e\u4fdd\u6587\u672c\u975e\u7a7a\n                embeddings = embedder.encode(document)   \n                embeddings_list.append(embeddings)\n                print(\"document:\", document)\n                print(f\"embeddings:{len(embeddings)},\", embeddings)\n                documents.append(document)\n                document = \"\"\n        document += line.strip() + \" \"\n    document = document.strip()\n    if document:  # \u5904\u7406\u6587\u4ef6\u6700\u540e\u7684\u5185\u5bb9\n        embeddings = embedder.encode(document)\n        embeddings_list.append(embeddings)\n        documents.append(document)<\/pre><\/div>\n\n\n\n<p>\u91cc\u9762\u7684 1000\uff0c \u53ef\u4ee5\u4fee\u6539\u4e3a\u66f4\u5927\u7684\u503c\uff0c\u6bd4\u598210000\uff0c\u8fd9\u6837\u5185\u5bb9\u53ef\u80fd\u66f4\u7cbe\u786e\uff0c\u5f53\u7136\u53d6\u5f97\u5185\u5bb9\u7684\u5927\u5c0f\uff0c\u53ef\u80fd\u5f71\u54cd\u63a8\u7406\u7684\u8f93\u5165\u957f\u5ea6\u3002<\/p>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>6. \u521b\u5efaFAISS\u7d22\u5f15\uff0c\u5e76\u4ea7\u751f\u63d0\u95ee<\/strong><\/h2>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:python decode:true \" ># \u521b\u5efaFAISS\u7d22\u5f15\nif embeddings_list:\n    embeddings_array = np.vstack(embeddings_list)\n    index = faiss.IndexFlatL2(embeddings_array.shape[1])\n    index.add(embeddings_array.astype('float32'))\n    \n    # \u7528\u6237\u95ee\u9898\u5904\u7406\u4e0e\u63a8\u7406\n    question = \"What is the theme of the document? \"\n    #question = \"\u8fd9\u4efd\u6587\u6863\u7684\u4e3b\u9898\u662f\u4ec0\u4e48\uff1f\"\n    query_embedding = embedder.encode([question])[0].astype('float32')\n    \n    # \u68c0\u7d22\u6700\u76f8\u5173\u7684\u51e0\u4e2a\u6587\u6863\u6bb5\u843d\n    combined_segments = \"\"\n    k = 3  # \u4f60\u5e0c\u671b\u68c0\u7d22\u7684\u76f8\u5173\u6587\u6863\u6570\u91cf\n    D, I = index.search(np.array([query_embedding]), k=k)\n    print(\"Top\", k, \"most relevant document segments:\")\n    for idx, segment_index in enumerate(I[0]):\n        most_relevant_segment = documents[segment_index]\n        print(f\"{idx+1}: {most_relevant_segment}\\n\")\n        combined_segments += \" \" + most_relevant_segment\n    \n    prompt = combined_segments + \"\\n\\n###\\n\\n\" + question + \"\\n\\n\u7528\u4e2d\u6587\u56de\u7b54\"\n    messages = [\n        {\"role\": \"system\", \"content\": \"You are a helpful assistant.\"},\n        {\"role\": \"user\", \"content\": prompt}\n    ]\n    text = tokenizer.apply_chat_template(\n        messages,\n        tokenize=False,\n        add_generation_prompt=True\n    )\n    model_inputs = tokenizer([text], return_tensors=\"pt\").to(device)\n    \n    generated_ids = model.generate(\n        model_inputs.input_ids,\n        max_new_tokens=512\n    )\n    generated_ids = [\n        output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)\n    ]\n    \n    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]    \n    print(\"Answer to the question:\", response)\nelse:\n    print(\"No embeddings found. Please check your data.\")<\/pre><\/div>\n\n\n\n<p>\u5bf9\u4e8e\u82f1\u6587\u6587\u6863\uff0c\u5e94\u8be5\u4ee5\u82f1\u6587\u63d0\u793a\uff0c\u4e2d\u6587\u6587\u6863\u4ee5\u4e2d\u6587\u63d0\u793a\u3002<\/p>\n\n\n\n<p>\u5982\u679c\u5e94\u4e2d\u6587\u63d0\u793a\u82f1\u6587\u7684\u6587\u6863\uff0c\u9700\u8981\u628a\u5d4c\u5165\u4fee\u6539\u4e00\u4e0b<\/p>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:python decode:true \" >embedder = SentenceTransformer('bert-base-multilingual-cased')<\/pre><\/div>\n\n\n\n<p>\u5bf9\u4e8e\u82f1\u6587\u6587\u6863\uff0c\u5982\u679c\u9700\u8981\u4e2d\u6587\u56de\u7b54\uff0c\u53ef\u4ee5\u5728\u540e\u9762\u5f3a\u5236\u8981\u6c42\uff1a<\/p>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:python decode:true \">prompt = combined_segments + \"\\n\\n###\\n\\n\" + question + \"\\n\\n\u7528\u4e2d\u6587\u56de\u7b54\"<\/pre><\/div>\n\n\n\n<p>\u4f60\u5e0c\u671b\u68c0\u7d22\u7684\u76f8\u5173\u6587\u6863\u6570\u91cf\uff0c\u901a\u8fc7\u8bbe\u7f6e k \u6765\u5b9e\u73b0<\/p>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:python decode:true \">k = 3  # \u4f60\u5e0c\u671b\u68c0\u7d22\u7684\u76f8\u5173\u6587\u6863\u6570\u91cf<\/pre><\/div>\n\n\n\n<p>\u4e0b\u9762\u662f\u8f93\u51fa\u7684\u7ed3\u679c\uff1a<\/p>\n\n\n\n<div class=\"wp-block-urvanov-syntax-highlighter-code-block\"><pre class=\"lang:sh decode:true \">Top 3 most relevant document segments:\n1: Over the next several years I wrote lots of essays about all kinds of different topics. O'Reilly reprinted a collection of them as a book, called Hackers &amp; Painters after one of the essays in it. I also worked on spam filters, and did some more painting. I used to have dinners for a group of friends every thursday night, which taught me how to cook for groups. And I bought another building in Cambridge, a former candy factory (and later, twas said, porn studio), to use as an office.\n\n2: AI was in the air in the mid 1980s, but there were two things especially that made me want to work on it: a novel by Heinlein called The Moon is a Harsh Mistress, which featured an intelligent computer called Mike, and a PBS documentary that showed Terry Winograd using SHRDLU. I haven't tried rereading The Moon is a Harsh Mistress, so I don't know how well it has aged, but when I read it I was drawn entirely into its world. It seemed only a matter of time before we'd have Mike, and when I saw Winograd using SHRDLU, it seemed like that time would be a few years at most. All you had to do was teach SHRDLU more words.\n\n3: We invited about 20 of the 225 groups to interview in person, and from those we picked 8 to fund. They were an impressive group. That first batch included reddit, Justin Kan and Emmett Shear, who went on to found Twitch, Aaron Swartz, who had already helped write the RSS spec and would a few years later become a martyr for open access, and Sam Altman, who would later become the second president of YC. I don't think it was entirely luck that the first batch was so good. You had to be pretty bold to sign up for a weird thing like the Summer Founders Program instead of a summer job at a legit place like Microsoft or Goldman Sachs.\n\nAnswer to the question: \u8fd9\u7bc7\u6587\u6863\u7684\u4e3b\u9898\u662f\u5173\u4e8e\u4f5c\u8005\u5728\u4e0d\u540c\u65f6\u671f\u64b0\u5199\u548c\u7814\u7a76\u4e0d\u540c\u4e3b\u9898\uff0c\u5305\u62ec\u4ed6\u5728\u827e\u8428\u514b\u00b7\u963f\u897f\u83ab\u592b\u7684\u300a\u6708\u7403\u4e0e\u6df1\u6e0a\u300b\uff08The Moon is a Harsh Mistress\uff09\u4e2d\u521b\u9020\u7684\u667a\u80fd\u8ba1\u7b97\u673a\u8fc8\u514b\u4ee5\u53ca\u4ed6\u7684\u7f16\u7a0b\u8bed\u8a00SHRDLU\u3002\u6b64\u5916\uff0c\u4ed6\u8fd8\u901a\u8fc7\u6559\u6388SHRDLU\u66f4\u591a\u7684\u5355\u8bcd\u6765\u53d1\u5c55AI\u6280\u672f\uff0c\u5e76\u53c2\u4e0e\u4e86\u7b2c\u4e00\u5c4aSummer Founders Program\u7684\u53d1\u8d77\u4eba\u4f1a\u8bae\uff0c\u5176\u4e2d\u5305\u62ecReddit\u3001Justin Kan\u3001\u57c3\u7c73\u7279\u00b7\u8c22\u62c9\uff08Emmet Shear\uff09\u3001Twitch\u521b\u59cb\u4ebaAaron Swartz\u3001\u4ed6\u4e4b\u524d\u5e2e\u52a9\u7f16\u5199RSS\u534f\u8bae\u5e76\u5c06\u5728\u51e0\u5e74\u540e\u6210\u4e3a\u5f00\u653e\u83b7\u53d6\u7684\u5148\u9a71\uff0c\u4ee5\u53caSam Altman\uff0c\u540e\u6765\u6210\u4e3aYC\u7684\u7b2c\u4e8c\u4efb\u6821\u957f\u3002\n\n\u5c3d\u7ba1\u6700\u521d\u7684\u96c6\u9526\u5305\u62ec\u4e86\u4e00\u4e9b\u6210\u529f\u8005\uff0c\u5982Reddit\u3001Justin Kan\u548c\u57c3\u7c73\u7279\u00b7\u8c22\u62c9\u7b49\uff0c\u4f46\u4f5c\u8005\u8ba4\u4e3a\u4ed6\u4eec\u7684\u6210\u529f\u5e76\u4e0d\u662f\u5b8c\u5168\u5076\u7136\u7684\uff0c\u800c\u662f\u6e90\u4e8e\u4ed6\u5bf9AI\u6280\u672f\u7684\u52c7\u6c14\u548c\u51b3\u5fc3\u3002\u4ed6\u660e\u786e\u8868\u793a\u8981\u53c2\u52a0\u4e00\u4e2a\u50cf\u662f\u5fae\u8f6f\u6216\u9ad8\u76db\u8fd9\u6837\u7684\u5408\u6cd5\u5de5\u4f5c\u800c\u975e\u4f20\u7edf\u7684\u6691\u671f\u5b9e\u4e60\u9879\u76ee\uff0c\u8fd9\u4e3a\u4ed6\u8fdb\u5165AI\u9886\u57df\u94fa\u5e73\u4e86\u9053\u8def\u3002\u6700\u7ec8\uff0c\u4f5c\u8005\u9009\u62e9\u521b\u5efaSHRDLU\u4f5c\u4e3a\u4eba\u5de5\u667a\u80fd\u7684\u7814\u7a76\u5de5\u5177\uff0c\u5e76\u9080\u8bf7\u5927\u7ea6225\u4e2a\u6765\u81ea\u5404\u79cd\u884c\u4e1a\u7684\u5c0f\u7ec4\u8fdb\u884c\u9762\u5bf9\u9762\u8bbf\u8c08\uff0c\u4ece\u4e2d\u6311\u9009\u51fa8\u4e2a\u8d44\u52a9\u5c0f\u7ec4\u3002\u8fd9\u4e9b\u5c0f\u7ec4\u5305\u62ec\u77e5\u540d\u7684\u6280\u672f\u516c\u53f8\uff0c\u5982Twitch\u3001AMD\u548c\u96c5\u864e\uff0c\u4ed6\u4eec\u5728\u968f\u540e\u7684\u53d1\u5c55\u5386\u7a0b\u4e2d\u626e\u6f14\u4e86\u91cd\u8981\u7684\u89d2\u8272\u3002\u56e0\u6b64\uff0c\u8fd9\u7bc7\u6587\u6863\u7684\u4e3b\u9898\u53ef\u4ee5\u6982\u62ec\u4e3a\u201c\u4f5c\u8005\u5728AI\u9886\u57df\u7684\u63a2\u7d22\u548c\u63a8\u52a8\uff0c\u4ee5\u53ca\u5176\u521b\u5efa\u548c\u5f71\u54cd\u4eba\u5de5\u667a\u80fd\u6280\u672f\u7684\u7b56\u7565\u201d\u3002<\/pre><\/div>\n","protected":false},"excerpt":{"rendered":"<p>\u672c\u6559\u7a0b\u4f7f\u7528\u4e86Qwen1.5-0.5B-Chat \u4f5c\u4e3a\u5d4c\u5165\u7684\u5c0f\u6a21\u578b \u4f7f\u7528 Qwen1.5-1.8B-Chat \u4f5c [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"set","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[444,443,442],"tags":[422],"class_list":["post-3622","post","type-post","status-publish","format-standard","hentry","category-ai","category-llm","category-llms","tag-rag"],"views":2584,"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/posts\/3622","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=3622"}],"version-history":[{"count":14,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/posts\/3622\/revisions"}],"predecessor-version":[{"id":3640,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=\/wp\/v2\/posts\/3622\/revisions\/3640"}],"wp:attachment":[{"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=3622"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=3622"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.aqwu.net\/wp\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=3622"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}