{"id":19757,"date":"2025-12-09T20:37:36","date_gmt":"2025-12-09T12:37:36","guid":{"rendered":"https:\/\/92it.top\/?p=19757"},"modified":"2025-12-09T20:37:36","modified_gmt":"2025-12-09T12:37:36","slug":"ai%e9%a2%86%e5%9f%9f%e7%9a%84%e5%85%b3%e9%94%ae%e6%8a%80%e6%9c%af-embedding%ef%bc%88%e5%b5%8c%e5%85%a5%ef%bc%89","status":"publish","type":"post","link":"https:\/\/92it.top\/?p=19757","title":{"rendered":"AI\u9886\u57df\u7684\u5173\u952e\u6280\u672f\u2014\u2014Embedding\uff08\u5d4c\u5165\uff09"},"content":{"rendered":"\n

\u8f6c\u8f7d\uff1ahttps:\/\/www.zhihu.com\/question\/561078975\/answer\/1956297309785658762<\/a><\/p>\n\n\n\n

\u4ec0\u4e48\u662fEmbedding\uff1f\ud83d\udd16<\/strong><\/p>\n\n\n\n

\n\n\n\n

\u7b80\u5355\u6765\u8bf4\uff0cEmbedding\u5c31\u662f\u628a\u6587\u5b57\u3001\u56fe\u7247\u6216\u5176\u4ed6\u6570\u636e\u8f6c\u6362\u6210\u6570\u5b57\u7684\u5f62\u5f0f\uff0c\u8ba9\u8ba1\u7b97\u673a\u80fd\u591f\u7406\u89e3\u548c\u5904\u7406<\/strong>\u3002\u5c31\u50cf\u6211\u4eec\u7ed9\u6bcf\u4e2a\u5355\u8bcd\u5206\u914d\u4e00\u4e2a\u300c\u6570\u5b57\u8eab\u4efd\u8bc1\u300d\uff0c\u4f46\u8fd9\u4e2a\u8eab\u4efd\u8bc1\u4e0d\u662f\u7b80\u5355\u7684\u7f16\u53f7\uff0c\u800c\u662f\u5305\u542b\u4e30\u5bcc\u4fe1\u606f\u7684\u300c\u591a\u7ef4\u7279\u5f81\u5411\u91cf\u300d\u3002<\/p>\n\n\n\n

Embedding \u6a21\u578b\u7684\u6838\u5fc3\u4ef7\u503c\u5728\u4e8e\u5c06\u975e\u7ed3\u6784\u5316\u6587\u672c\u8f6c\u5316\u4e3a\u6570\u503c\u5411\u91cf\uff0c\u89e3\u51b3\u8bed\u4e49\u7406\u89e3\u4e0e\u8ba1\u7b97\u6548\u7387\u95ee\u9898.<\/p>\n\n\n\n

\u5148\u4e3e\u4e00\u4e2a\u751f\u52a8\u7684\u6817\u5b50<\/strong><\/p>\n\n\n\n

\u60f3\u8c61\u4e00\u4e0b\uff0c\u6211\u4eec\u8981\u7ed9\u4e16\u754c\u4e0a\u6240\u6709\u98df\u7269\u7f16\u7801\uff1a<\/p>\n\n\n\n

\u6c49\u5821\uff1a[-0.2, 0.8, 0.3]\n\n\u6c99\u62c9\uff1a[0.5, -0.3, 0.1]\n\n\u51b0\u6dc7\u6dcb\uff1a[0.1, 0.9, -0.4]<\/pre>\n\n\n\n
\u6570\u5b57\u7684\u6b63\u8d1f\u548c\u5927\u5c0f\u4ee3\u8868\u4e86\u98df\u7269\u7684\u7279\u5f81\uff08\u5982\u751c\u5ea6\u3001\u5065\u5eb7\u7a0b\u5ea6\u3001\u6e29\u5ea6\uff09\u3002\u8fd9\u6837\u8ba1\u7b97\u673a\u5c31\u80fd\u901a\u8fc7\u6570\u5b57\u8ba1\u7b97\u53d1\u73b0\u300c\u6c49\u5821\u548c\u51b0\u6dc7\u6dcb\u90fd\u662f\u9ad8\u70ed\u91cf\u98df\u7269\u300d\u3002<\/p>\n\n\n
\n
\"\"<\/figure><\/div>\n\n\n
Embedding\u5c31\u662f\u5927\u6a21\u578b\u81ea\u5df1\u7684\u8bed\u8a00\uff0c\u4efb\u4f55\u9700\u8981\u8ddf\u5927\u6a21\u578b\u6c9f\u901a\u7684\u6587\u5b57\u3001\u56fe\u50cf\u3001\u89c6\u9891\u90fd\u9700\u8981\u8f6c\u6362\u4e3a\u5927\u6a21\u578b\u6240\u80fd\u7406\u89e3\u7684\u8bed\u8a00\uff1aEmbedding\uff0c\u5b83\u624d\u80fd\u5904\u7406\u3002\u5904\u7406\u5b8c\u6210\u540e\uff0c\u5b83\u518d\u7ffb\u8bd1\u6210\u4eba\u7c7b\u80fd\u7406\u89e3\u7684\u6587\u5b57\u3001\u56fe\u50cf\u7b49\u3002\u8fd9\u4e5f\u662f\u5927\u6a21\u578b\u6700\u5f3a\u5927\u7684\u6838\u5fc3\u80fd\u529b\u4e4b\u4e00\uff0c\u591a\u6a21\u6001\u5904\u7406\u80fd\u529b.<\/p>\n\n\n\n
\u3000\u3000<\/p>\n\n\n\n
Embedding\u7684\u4e00\u4e9b\u6848\u4f8b \ud83d\udd16<\/strong><\/p>\n\n\n\n
\n\n\n\n
\ud83d\udd39<\/strong><\/strong>\u8bed\u4e49\u7f16\u7801\uff1a\u901a\u8fc7\u5411\u91cf\u7a7a\u95f4\u6355\u6349\u4e0a\u4e0b\u6587\u5173\u8054\uff0c\u533a\u5206\u591a\u4e49\u8bcd\u3001\u540c\u4e49\u8bcd\uff1b\u5982\uff1a<\/p>\n\n\n\n
\u95ee\u9898\uff1a\u7528\u6237\u641c\u7d22\u201c\u82f9\u679c\u201d\uff0c\u9700\u533a\u5206\u201c\u6c34\u679c\u201d\u8fd8\u662f\u201c\u54c1\u724c\u201d\u3002<\/p>\n\n\n\n
Embedding \u4f5c\u7528\uff1a<\/p>\n\n\n\n
\"\u82f9\u679c\u624b\u673a\"\u7684\u5411\u91cf \u4f1a\u63a5\u8fd1 \"iPhone\"\u667a\u80fd\u624b\u673a\uff1b\n\n\u800c\"\u7ea2\u82f9\u679c\"\u7684\u5411\u91cf \u4f1a\u63a5\u8fd1 \"\u6c34\u679c\" \"\u9999\u8549\" \"\u7ef4\u751f\u7d20\"\u3002<\/pre>\n\n\n\n
\u7ed3\u679c\uff1a\u641c\u7d22\u201c\u82f9\u679c\u201d\u65f6\uff0c\u4f18\u5148\u5c55\u793a\u624b\u673a\u6216\u6c34\u679c\uff0c\u53d6\u51b3\u4e8e\u7528\u6237\u5386\u53f2\u884c\u4e3a\uff08\u5982\u70b9\u51fb\u7535\u5b50\u4ea7\u54c1\uff09\u3002<\/p>\n\n\n\n
\u3000\u3000<\/p>\n\n\n\n
\ud83d\udd39<\/strong><\/strong>\u9ad8\u6548\u68c0\u7d22\uff1a\u652f\u6301\u8fd1\u4f3c\u6700\u8fd1\u90bb\uff08ANN\uff09\u7b97\u6cd5\uff0c\u964d\u4f4e\u6d77\u91cf\u6570\u636e\u5339\u914d\u590d\u6742\u5ea6\uff1b<\/p>\n\n\n\n
\u573a\u666f\uff1a\u67d0\u7535\u5546\u5e73\u53f0\u6709 1 \u4ebf\u5546\u54c1\u63cf\u8ff0\uff0c\u9700\u5b9e\u65f6\u5339\u914d\u7528\u6237\u67e5\u8be2\u201c\u9002\u5408\u9732\u8425\u7684\u8f7b\u4fbf\u5e10\u7bf7\u201d\u3002<\/p>\n\n\n\n
\u4f20\u7edf\u65b9\u6cd5\uff1a\u5173\u952e\u8bcd\u5339\u914d\u201c\u9732\u8425+\u8f7b\u4fbf+\u5e10\u7bf7\u201d\uff0c\u53ef\u80fd\u6f0f\u6389\u201c\u6237\u5916\u8d85\u8584\u906e\u9633\u7bf7\u201d\u3002<\/p>\n\n\n\n
Embedding \u65b9\u6848\uff1a<\/p>\n\n\n\n
\u5c06\u67e5\u8be2\u548c\u5546\u54c1\u63cf\u8ff0\u8f6c\u4e3a\u5411\u91cf\uff1b\n\n\u4f7f\u7528 ANN \u5e93\uff08\u5982 FAISS \uff09\u5728\u6beb\u79d2\u7ea7\u8fd4\u56de<\/pre>\n\n\n\n
Top100 \u76f8\u5173\u5546\u54c1\uff0c\u8986\u76d6\u8bed\u4e49\u76f8\u4f3c\u4f46\u5173\u952e\u8bcd\u4e0d\u5339\u914d\u7684\u7ed3\u679c\u3002<\/p>\n\n\n\n
\u3000\u3000<\/p>\n\n\n\n
\u00a0\ud83d\udd39AI \u57fa\u7840\u8bbe\u65bd\uff1a\u652f\u6491 RAG \u3001\u591a\u6a21\u6001\u641c\u7d22\u3001\u8fc1\u79fb\u5b66\u4e60\u7b49\u4efb\u52a1\uff0c\u66ff\u4ee3\u4f20\u7edf\u5173\u952e\u8bcd\u5339\u914d\u4e0e\u4eba\u5de5\u89c4\u5219\u3002<\/p>\n\n\n\n
\u573a\u666f\uff1a<\/strong>\u5ba2\u670d\u673a\u5668\u4eba\u56de\u7b54\u201c\u5982\u4f55\u6e05\u6d01\u5e10\u7bf7\u4e0a\u7684\u6c61\u6e0d\uff1f\u201d<\/p>\n\n\n\n
\u6d41\u7a0b\uff1a<\/strong>\u7528 BGE-M3 \u5c06\u95ee\u9898\u7f16\u7801\u4e3a\u5411\u91cf\uff1b<\/p>\n\n\n\n
\u4ece\u5411\u91cf\u6570\u636e\u5e93\u68c0\u7d22\u300a\u6237\u5916\u7528\u54c1\u4fdd\u517b\u6307\u5357\u300b\u4e2d\u76f8\u5173\u6bb5\u843d\uff1b\n\n\u5c06\u68c0\u7d22\u7ed3\u679c\u8f93\u5165\u5927\u6a21\u578b\uff08\u5982DeepSeek-R1\uff09\uff0c\u751f\u6210\u6b65\u9aa4\u6e05\u6670\u7684\u56de\u7b54\u3002<\/pre>\n\n\n\n
\u4f18\u52bf\uff1a<\/strong>\u907f\u514d\u5927\u6a21\u578b\u865a\u6784\u7b54\u6848\uff0c\u63d0\u5347\u53ef\u4fe1\u5ea6\u3002<\/p>\n\n\n\n
\u3000\u3000<\/p>\n\n\n\n
\u91cd\u70b9\uff1a\u4ee3\u7801\u5b9e\u6218 \ud83d\udd16<\/strong><\/p>\n\n\n\n
\n\n\n\n
\ud83d\udd39\u597d\uff0c\u63a5\u4e0b\u6765\u76f4\u63a5\u4e0a\u4ee3\u7801\uff0c\u770bEmbedding\u5982\u4f55\u5de5\u4f5c<\/p>\n\n\n\n
# \u5bfc\u5165\u5fc5\u8981\u7684\u5e93\nfrom sentence_transformers import SentenceTransformer\n \n# \u52a0\u8f7d\u6a21\u578b\nmodel = SentenceTransformer('paraphrase-MiniLM-L6-v2')\n \n# \u5b9a\u4e49\u53e5\u5b50\u5217\u8868\nsentences = [\n    '\u6211\u559c\u6b22\u5403\u82f9\u679c',\n    '\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17',\n    '\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91',\n    \"\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c\"\n]\n \n# \u751f\u6210Embedding\u5411\u91cf\nembeddings = model.encode(sentences)\n \nprint(\"\u751f\u6210\u7684Embedding\u7684\u5411\u91cf\u7ef4\u5ea6\uff1a\", embeddings.shape)\n \n# \u8f93\u51fa\u5168\u90e8\n# print(\"Embedding\u5411\u91cf\u7684\u5177\u4f53\u5185\u5bb9\uff1a\", embeddings)\n \n# \u8f93\u51fa\u4e0a\u9762\u53e5\u5b50\u7684\u5411\u91cf\u524d10\u7ef4\nfor i, sentence in enumerate(sentences):\n    print(f\"'{sentence}': {embeddings[i][:10]} ...\")\n \n \n# \u8ba1\u7b97\u53e5\u5b50\u95f4\u7684\u76f8\u4f3c\u5ea6\nfrom sklearn.metrics.pairwise import cosine_similarity\n \n# \u8ba1\u7b97\u4f59\u5f26\u76f8\u4f3c\u5ea6\nsimilarities = cosine_similarity(embeddings)\n \nprint(\"\u53e5\u5b50\u76f8\u4f3c\u5ea6\u77e9\u9635\uff1a\")\nfor i in range(len(sentences)):\n    for j in range(len(sentences)):\n        if i <= j:  # \u907f\u514d\u91cd\u590d\u8f93\u51fa\n            sim = similarities[i][j]\n            print(f\"'{sentences[i]}' vs '{sentences[j]}': {sim:.3f}\")<\/pre>\n\n\n\n
\u8f93\u51fa\u5982\u4e0b\uff1a<\/p>\n\n\n\n
\u751f\u6210\u7684Embedding\u7684\u5411\u91cf\u7ef4\u5ea6\uff1a (4, 384)\n \n'\u6211\u559c\u6b22\u5403\u82f9\u679c': [-0.37381974  1.1482197  -0.10418612  0.09360497 -0.66173714  0.2802788\n1.6347399   0.14204119  0.8921655  -0.16619411] ...\n'\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17': [-0.2590392   0.6924811   0.18627527 -0.5357205  -0.07483669 -0.19914986\n0.8918247   0.05937493  0.6875115   0.24681543] ...\n'\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91': [-0.3749065   0.74014777  0.16411266 -0.21965387 -0.48084465 -0.22723575\n1.4115497   0.11568774  0.76047236 -0.31601837] ...\n'\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c': [-0.11292891  0.5817749   0.06361266  0.02144676 -0.511007    0.1269096\n1.2333398   0.3227354   1.0315524   0.00547485] ...\n \n\u53e5\u5b50\u76f8\u4f3c\u5ea6\u77e9\u9635\uff1a\n'\u6211\u559c\u6b22\u5403\u82f9\u679c' vs '\u6211\u559c\u6b22\u5403\u82f9\u679c': 1.000\n'\u6211\u559c\u6b22\u5403\u82f9\u679c' vs '\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17': 0.445\n'\u6211\u559c\u6b22\u5403\u82f9\u679c' vs '\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91': 0.782\n'\u6211\u559c\u6b22\u5403\u82f9\u679c' vs '\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c': 0.823\n'\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17' vs '\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17': 1.000\n'\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17' vs '\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91': 0.642\n'\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17' vs '\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c': 0.544\n'\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91' vs '\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91': 1.000\n'\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91' vs '\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c': 0.866\n'\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c' vs '\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c': 1.000<\/pre>\n\n\n\n
\u6ce8\uff1a\u4e0a\u9762\u7684\u4ee3\u7801\u4f1a\u9ed8\u8ba4\u4ecehugging face\u5b98\u65b9\u4ed3\u5e93\u4e0b\u8f7d\u6a21\u578b\uff0c\u9700\u8981\u7ffb\u5899\uff01\uff01\uff01\uff01\uff01<\/p>\n\n\n\n
\u3000\u3000<\/p>\n\n\n\n
\ud83d\udd39\u63a5\u4e0b\u6765\u518d\u7528ModelScope\u7684\u65b9\u5f0f\u5b9e\u73b0\u4e00\u904d\uff1a\uff08ModelScope\u963f\u91cc\u4e91\u63d0\u4f9b\u7684\u4e00\u4e2a\u4e2d\u6587\u7248\u7684HuggingFace\uff09<\/p>\n\n\n\n
from langchain_community.embeddings import ModelScopeEmbeddings\n \n# \u521d\u59cb\u5316\u53e5\u5b50\u5d4c\u5165pipeline\nmodel_id = \"damo\/nlp_corom_sentence-embedding_chinese-base\"\nembedding_model = ModelScopeEmbeddings(model_id=model_id)\n \n# \u5b9a\u4e49\u53e5\u5b50\u5217\u8868\nsentences = [\n    '\u6211\u559c\u6b22\u5403\u82f9\u679c',\n    '\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17',\n    '\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91',\n    \"\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c\"\n]\n \nembeddings = embedding_model.embed_documents(sentences)\n \n# \u8f93\u51fa\u5168\u90e8\n# print(\"Embedding\u5411\u91cf\u7684\u5177\u4f53\u5185\u5bb9\uff1a\", embeddings)\n \n# \u8f93\u51fa\u4e0a\u9762\u53e5\u5b50\u7684\u5411\u91cf\u524d10\u7ef4\nfor i, sentence in enumerate(sentences):\n    print(f\"'{sentence}': {embeddings[i][:10]} ...\")\n \n \nfrom sklearn.metrics.pairwise import cosine_similarity\n \n# \u8ba1\u7b97\u4f59\u5f26\u76f8\u4f3c\u5ea6\u77e9\u9635\nsimilarity_matrix = cosine_similarity(embeddings)\n \nprint(\"\u53e5\u5b50\u4e4b\u95f4\u7684\u4f59\u5f26\u76f8\u4f3c\u5ea6\u77e9\u9635:\")\nfor i in range(len(sentences)):\n    for j in range(len(sentences)):\n        if i <= j:  # \u907f\u514d\u91cd\u590d\u8f93\u51fa\n            print(f\"'{sentences[i]}' vs '{sentences[j]}' \u7684\u76f8\u4f3c\u5ea6: {similarity_matrix[i][j]:.4f}\")<\/pre>\n\n\n\n
\u8f93\u51fa\u5982\u4e0b\uff1a<\/p>\n\n\n\n
'\u6211\u559c\u6b22\u5403\u82f9\u679c': [0.32604819536209106, -0.24579137563705444, 0.14939190447330475, -0.18595966696739197, -0.2441755086183548, -0.5086973309516907, -0.4876851439476013, -0.01850845105946064, 0.44076400995254517, -0.019114553928375244] ...\n'\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17': [0.1550225019454956, -0.2128225564956665, 0.2851581275463104, -0.8280717730522156, 0.5369096994400024, -0.7610117793083191, -0.14223024249076843, -0.8344448208808899, -0.31277164816856384, -0.0666898638010025] ...\n'\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91': [0.6082981824874878, 0.3789174258708954, 0.004537384957075119, -0.027960695326328278, 0.46887874603271484, -0.5822398662567139, -0.4064663350582123, -1.11110520362854, 0.24349763989448547, 0.1817186176776886] ...\n'\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c': [0.38078010082244873, -0.6754599213600159, 0.1290031373500824, -0.14763924479484558, -0.011951431632041931, -0.23197220265865326, -0.3167392313480377, -1.2026702165603638, -0.014405147172510624, 0.2859038710594177] ...\n \n\u53e5\u5b50\u4e4b\u95f4\u7684\u4f59\u5f26\u76f8\u4f3c\u5ea6\u77e9\u9635:\n'\u6211\u559c\u6b22\u5403\u82f9\u679c' vs '\u6211\u559c\u6b22\u5403\u82f9\u679c' \u7684\u76f8\u4f3c\u5ea6: 1.0000\n'\u6211\u559c\u6b22\u5403\u82f9\u679c' vs '\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17' \u7684\u76f8\u4f3c\u5ea6: 0.4424\n'\u6211\u559c\u6b22\u5403\u82f9\u679c' vs '\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91' \u7684\u76f8\u4f3c\u5ea6: 0.4736\n'\u6211\u559c\u6b22\u5403\u82f9\u679c' vs '\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c' \u7684\u76f8\u4f3c\u5ea6: 0.6298\n'\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17' vs '\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17' \u7684\u76f8\u4f3c\u5ea6: 1.0000\n'\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17' vs '\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91' \u7684\u76f8\u4f3c\u5ea6: 0.4052\n'\u82f9\u679c\u516c\u53f89\u6708\u8981\u53d1\u5e03\u65b0\u624b\u673aiPhone 17' vs '\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c' \u7684\u76f8\u4f3c\u5ea6: 0.4137\n'\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91' vs '\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91' \u7684\u76f8\u4f3c\u5ea6: 1.0000\n'\u4eca\u5929\u7684\u5929\u6c14\u771f\u597d\uff0c\u84dd\u5929\u767d\u4e91' vs '\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c' \u7684\u76f8\u4f3c\u5ea6: 0.4308\n'\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c' vs '\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c' \u7684\u76f8\u4f3c\u5ea6: 1.0000<\/pre>\n\n\n\n
\u4e24\u79cd\u5b9e\u73b0\u65b9\u5f0f\u5927\u4f53\u4e0a\u7ed3\u679c\u662f\u4e00\u6837\u7684\uff0c\u56e0\u4e3a\u662f\u4e0d\u540c\u7684\u6a21\u578b\uff0c\u6240\u4ee5\u5411\u91cf\u503c\u4f1a\u6709\u5dee\u5f02\u3002\uff08\u4e0a\u9762\u4e24\u79cd\u5b9e\u73b0\u65b9\u5f0f\u4e09\u65b9\u4f9d\u8d56\u53ef\u80fd\u4f1a\u5b58\u5728\u51b2\u7a81\uff0c\u5efa\u8bae\u4f7f\u7528\u4e0d\u540c\u7684\u73af\u5883\u5c1d\u8bd5\u5594\uff01\uff09<\/p>\n\n\n\n
\u662f\u4e0d\u662f\u5f88\u795e\u5947\uff01\u8ba1\u7b97\u673a\u901a\u8fc7Embedding\u53d1\u73b0\uff1a<\/p>\n\n\n\n
    \n
  • 1\u3001\u300c\u6211\u559c\u6b22\u5403\u82f9\u679c\u300d\u548c\u300c\u9999\u8549\u662f\u4e00\u79cd\u70ed\u5e26\u6c34\u679c\u300d\u76f8\u4f3c\u5ea6\u5f88\u9ad8\uff08\u90fd\u662f\u6c34\u679c\uff09<\/li>\n\n\n\n
  • 2\u3001\u4f46\u4e0e\u300c\u82f9\u679c\u516c\u53f8\u300d\u76f8\u4f3c\u5ea6\u5f88\u4f4e\uff08\u867d\u7136\u90fd\u6709”\u82f9\u679c”\u4e8c\u5b57\uff09<\/li>\n\n\n\n
  • 3\u3001\u548c\u300c\u5929\u6c14\u300d\u57fa\u672c\u4e0d\u76f8\u5173<\/li>\n<\/ul>\n\n\n\n
    \u8fd9\u5c31\u662fEmbedding\u7684\u9b54\u529b\u2014\u2014\u7406\u89e3\u8bed\u4e49\u800c\u4e0d\u4ec5\u4ec5\u662f\u5b57\u9762\uff01<\/p>\n\n\n\n
    \u3000\u3000<\/p>\n\n\n\n
    \ud83d\udd39\u4e3a\u4ec0\u4e48Embedding\u5982\u6b64\u91cd\u8981\uff1f<\/p>\n\n\n\n
      \n
    • 1\u3001\u8ba9\u8ba1\u7b97\u673a\u7406\u89e3\u8bed\u8a00\uff1a\u4e0d\u518d\u662f\u7b80\u5355\u7684\u5173\u952e\u8bcd\u5339\u914d\uff0c\u800c\u662f\u771f\u6b63\u7684\u8bed\u4e49\u7406\u89e3<\/li>\n\n\n\n
    • 2\u3001\u8d4b\u80fd\u63a8\u8350\u7cfb\u7edf\uff1aNetflix\u3001\u6dd8\u5b9d\u90fd\u5728\u7528Embedding\u7406\u89e3\u4f60\u7684\u559c\u597d<\/li>\n\n\n\n
    • 3\u3001\u9a71\u52a8\u641c\u7d22\u5f15\u64ce\uff1aGoogle\u641c\u7d22\u80cc\u540e\u7684\u6838\u5fc3\u79d1\u6280\u4e4b\u4e00<\/li>\n\n\n\n
    • 4\u3001\u804a\u5929\u673a\u5668\u4eba\uff1aChatGPT\u7b49AI\u5bf9\u8bdd\u7cfb\u7edf\u7684\u57fa\u77f3<\/li>\n<\/ul>\n\n\n\n
      \u3000\u3000<\/p>\n\n\n\n
      \u76ee\u524d\u6d41\u884c\u7684Embedding\u6a21\u578b \ud83d\udd16<\/strong><\/p>\n\n\n\n
      \n\n\n
      \n
      \"\"<\/figure><\/div>\n\n
      \n
      \"\"<\/figure><\/div>\n\n\n
      \u8bf4\u5728\u6700\u540e\uff1aEmbedding\u5c31\u50cf\u662f\u4e00\u5ea7\u6865\u6881\uff0c\u8fde\u63a5\u4e86\u4eba\u7c7b\u8bed\u8a00\u548c\u8ba1\u7b97\u673a\u6570\u5b57\u4e16\u754c\u3002\u5b83\u901a\u8fc7\u5c06\u6587\u5b57\u8f6c\u6362\u4e3a\u5bcc\u542b\u8bed\u4e49\u7684\u6570\u5b57\u5411\u91cf\uff0c\u8ba9\u673a\u5668\u80fd\u591f\u771f\u6b63\u300c\u7406\u89e3\u300d\u6211\u4eec\u7684\u8bed\u8a00\u3002\u4e0b\u6b21\u5f53\u4f60\u4f7f\u7528\u667a\u80fd\u641c\u7d22\u3001\u63a8\u8350\u7cfb\u7edf\u6216\u4e0eAI\u5bf9\u8bdd\u65f6\uff0c\u5c31\u77e5\u9053\u80cc\u540e\u662fEmbedding\u6280\u672f\u5728\u9ed8\u9ed8\u5de5\u4f5c\u5566\uff01<\/p>\n","protected":false},"excerpt":{"rendered":"
      \u8f6c\u8f7d\uff1ahttps:\/\/www.zhihu.com\/question\/561078975\/answer\/1956 […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[48],"tags":[],"_links":{"self":[{"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/posts\/19757"}],"collection":[{"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/92it.top\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=19757"}],"version-history":[{"count":6,"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/posts\/19757\/revisions"}],"predecessor-version":[{"id":19767,"href":"https:\/\/92it.top\/index.php?rest_route=\/wp\/v2\/posts\/19757\/revisions\/19767"}],"wp:attachment":[{"href":"https:\/\/92it.top\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=19757"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/92it.top\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=19757"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/92it.top\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=19757"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}