betterdocs
domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init
action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /data/user/htdocs/wp-includes/functions.php on line 6114jnews-view-counter
domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init
action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /data/user/htdocs/wp-includes/functions.php on line 6114wp-statistics
domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init
action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /data/user/htdocs/wp-includes/functions.php on line 6114wpdiscuz
domain was triggered too early. This is usually an indicator for some code in the plugin or theme running too early. Translations should be loaded at the init
action or later. Please see Debugging in WordPress for more information. (This message was added in version 6.7.0.) in /data/user/htdocs/wp-includes/functions.php on line 6114jnews
域的翻译加载触发过早。这通常表示插件或主题中的某些代码运行过早。翻译应在 init
操作或之后加载。 请查阅调试 WordPress来获取更多信息。 (这个消息是在 6.7.0 版本添加的。) in /data/user/htdocs/wp-includes/functions.php on line 6114jnews-like
域的翻译加载触发过早。这通常表示插件或主题中的某些代码运行过早。翻译应在 init
操作或之后加载。 请查阅调试 WordPress来获取更多信息。 (这个消息是在 6.7.0 版本添加的。) in /data/user/htdocs/wp-includes/functions.php on line 6114\u8bad\u7ec3\u524d\u6cbf\u5927\u578b\u591a\u6a21\u6001\u6a21\u578b \uff08LMM\uff09 \u9700\u8981\u5927\u89c4\u6a21\u6570\u636e\u96c6\uff0c\u5176\u4e2d\u5305\u542b\u81ea\u7531\u5f62\u5f0f\u7684\u56fe\u50cf\u548c\u6587\u672c\u4ea4\u9519\u5e8f\u5217\u3002\u5c3d\u7ba1\u5f00\u6e90 LMM \u53d1\u5c55\u8fc5\u901f\uff0c\u4f46\u5f00\u6e90\u7684\u5927\u89c4\u6a21\u591a\u6a21\u6001\u4ea4\u9519\u6570\u636e\u96c6\u4ecd\u7136\u4e25\u91cd\u7f3a\u4e4f\u3002\u8fd9\u4e9b\u6570\u636e\u96c6\u7684\u91cd\u8981\u6027\u600e\u4e48\u5f3a\u8c03\u90fd\u4e0d\u4e3a\u8fc7\uff0c\u56e0\u4e3a\u5b83\u4eec\u4e3a\u521b\u5efa\u80fd\u591f\u7406\u89e3\u548c\u751f\u6210\u4e0d\u540c\u6a21\u6001\u5185\u5bb9\u7684\u5148\u8fdb\u4eba\u5de5\u667a\u80fd\u7cfb\u7edf\u5960\u5b9a\u4e86\u57fa\u7840\u3002\u5982\u679c\u6ca1\u6709\u8db3\u591f\u7684\u5168\u9762\u3001\u4ea4\u9519\u7684\u6570\u636e\u96c6\u4f9b\u5e94\uff0c\u5f00\u53d1\u66f4\u590d\u6742\u3001\u66f4\u6709\u80fd\u529b\u7684LMM\u7684\u6f5c\u529b\u5c31\u4f1a\u53d7\u5230\u4e25\u91cd\u963b\u788d\u3002\u8fd9\u4e9b\u6570\u636e\u96c6\u4f7f\u6a21\u578b\u80fd\u591f\u4ece\u5404\u79cd\u8f93\u5165\u4e2d\u5b66\u4e60\uff0c\u4f7f\u5b83\u4eec\u5728\u5404\u79cd\u5e94\u7528\u4e2d\u66f4\u52a0\u901a\u7528\u548c\u6709\u6548\u3002\u6b64\u5916\uff0c\u6b64\u7c7b\u6570\u636e\u96c6\u7684\u7a00\u7f3a\u6027\u5bf9\u5f00\u6e90\u793e\u533a\u6784\u6210\u4e86\u6311\u6218\uff0c\u5f00\u6e90\u793e\u533a\u4f9d\u9760\u5171\u4eab\u8d44\u6e90\u6765\u63a8\u52a8\u521b\u65b0\u548c\u534f\u4f5c\u3002<\/p>\n
\u8fd1\u5e74\u6765\uff0c\u5f00\u6e90 LMM \u53d6\u5f97\u4e86\u957f\u8db3\u7684\u8fdb\u6b65\uff0c\u4f46\u5176\u589e\u957f\u53d7\u5230\u5927\u89c4\u6a21\u4ea4\u9519\u6570\u636e\u96c6\u53ef\u7528\u6027\u6709\u9650\u7684\u963b\u788d\u3002\u4e3a\u4e86\u514b\u670d\u8fd9\u4e00\u969c\u788d\uff0c\u9700\u8981\u9f50\u5fc3\u534f\u529b\u5730\u7b56\u5212\u3001\u6ce8\u91ca\u548c\u53d1\u5e03\u66f4\u5168\u9762\u7684\u6570\u636e\u96c6\uff0c\u4ee5\u652f\u6301\u591a\u6a21\u6001\u6a21\u578b\u7684\u6301\u7eed\u5f00\u53d1\u548c\u5b8c\u5584\u3002\u6b64\u5916\uff0c\u8fd9\u4e9b\u6570\u636e\u96c6\u7684\u521b\u5efa\u548c\u4f20\u64ad\u6d89\u53ca\u514b\u670d\u4e00\u4e9b\u6280\u672f\u548c\u540e\u52e4\u969c\u788d\u3002\u6570\u636e\u6536\u96c6\u5fc5\u987b\u662f\u5e7f\u6cdb\u7684\uff0c\u5e76\u80fd\u4ee3\u8868\u90e8\u7f72LMMs\u7684\u5404\u79cd\u60c5\u51b5\u3002\u6ce8\u91ca\u9700\u8981\u4ed4\u7ec6\u8003\u8651\uff0c\u4ee5\u786e\u4fdd\u56fe\u50cf\u548c\u6587\u672c\u7684\u4ea4\u9519\u5e8f\u5217\u4ee5\u589e\u5f3a\u6a21\u578b\u5b66\u4e60\u80fd\u529b\u7684\u65b9\u5f0f\u5bf9\u9f50\u3002\u6b64\u5916\uff0c\u786e\u4fdd\u6570\u636e\u96c6\u662f\u5f00\u6e90\u7684\uff0c\u9700\u8981\u89e3\u51b3\u4e0e\u6570\u636e\u9690\u79c1\u548c\u4f7f\u7528\u6743\u76f8\u5173\u7684\u6cd5\u5f8b\u548c\u9053\u5fb7\u8003\u8651\u3002\u6269\u5927\u9ad8\u8d28\u91cf\u3001\u5927\u89c4\u6a21\u591a\u6a21\u6001\u4ea4\u9519\u6570\u636e\u96c6\u7684\u53ef\u7528\u6027\u5bf9\u4e8e\u4eba\u5de5\u667a\u80fd\u7814\u7a76\u548c\u5f00\u53d1\u7684\u672a\u6765\u81f3\u5173\u91cd\u8981\u3002\u901a\u8fc7\u89e3\u51b3\u5f53\u524d\u7684\u7a00\u7f3a\u6027\u95ee\u9898\uff0c\u4eba\u5de5\u667a\u80fd\u793e\u533a\u53ef\u4ee5\u4fc3\u8fdb\u66f4\u5927\u7684\u521b\u65b0\u548c\u534f\u4f5c\uff0c\u4ece\u800c\u521b\u5efa\u66f4\u5f3a\u5927\u3001\u66f4\u901a\u7528\u7684 LMM\uff0c\u80fd\u591f\u89e3\u51b3\u590d\u6742\u7684\u73b0\u5b9e\u4e16\u754c\u95ee\u9898\u3002<\/p>\n
\u5728\u6b64\u57fa\u7840\u4e0a\uff0cMINT-1T \u662f\u8fc4\u4eca\u4e3a\u6b62\u89c4\u6a21\u6700\u5927\u3001\u6700\u591a\u6837\u5316\u7684\u591a\u6a21\u6001\u4ea4\u9519\u5f00\u6e90\u6570\u636e\u96c6\u3002MINT-1T\uff1a\u6bd4\u73b0\u6709\u7684\u5f00\u6e90\u6570\u636e\u96c6\u5927 10 \u500d\uff0c\u5305\u62ec 1 \u4e07\u4ebf\u4e2a\u6587\u672c\u6807\u8bb0\u548c 34 \u4ebf\u5f20\u56fe\u50cf\u3002MINT-1T \u6570\u636e\u96c6\u8fd8\u5f15\u5165\u4e86\u4ece\u672a\u516c\u5f00\u7684\u6765\u6e90\uff0c\u4f8b\u5982 PDF \u6587\u4ef6\u3001ArXiv \u8bba\u6587\u3002\u7531\u4e8e\u591a\u6a21\u6001\u4ea4\u9519\u6570\u636e\u96c6\u4e0d\u5bb9\u6613\u6269\u5c55\uff0c\u56e0\u6b64 MINT-1T \u6570\u636e\u96c6\u5171\u4eab\u6570\u636e\u7ba1\u7406\u8fc7\u7a0b\u975e\u5e38\u91cd\u8981\uff0c\u8fd9\u6837\u5176\u4ed6\u4eba\u4e5f\u53ef\u4ee5\u5bf9\u6b64\u7c7b\u4fe1\u606f\u4e30\u5bcc\u7684\u53d8\u4f53\u8fdb\u884c\u5b9e\u9a8c\u3002MINT-1T\u6570\u636e\u96c6\u8bc1\u660e\u4e86\u5176\u65b9\u6cd5;\u5728MINT-1T\u4e0a\u8bad\u7ec3\u7684LM\u6a21\u578b\u4e0e\u4ee5\u524d\u6700\u5148\u8fdb\u7684\u65b9\u5c16\u58f3\u76f8\u6bd4\u5177\u6709\u7ade\u4e89\u529b\uff08\u5c3d\u7ba1\u6709\u4e9b\u4e0d\u540c\uff09\u3002<\/p>\n
<\/p>\n
MINT-1T<\/strong>\uff1a\u5177\u6709\u4e00\u4e07\u4ebf\u4e2a\u6807\u8bb0\u7684\u591a\u6a21\u6001\u6570\u636e\u96c6<\/strong><\/p>\n \u5927\u578b\u5f00\u6e90\u9884\u8bad\u7ec3\u6570\u636e\u96c6\u5bf9\u4e8e\u7814\u7a76\u754c\u63a2\u7d22\u6570\u636e\u5de5\u7a0b\u548c\u8bad\u7ec3\u900f\u660e\u7684\u5f00\u6e90\u6a21\u578b\u81f3\u5173\u91cd\u8981\u3002\u5728\u6587\u672c\u9886\u57df\uff0cC4 \u548c The Pile \u7b49\u65e9\u671f\u4f5c\u54c1\u5728\u4f7f\u793e\u533a\u80fd\u591f\u8bad\u7ec3\u7b2c\u4e00\u7ec4\u5f00\u6e90\u5927\u578b\u8bed\u8a00\u6a21\u578b\uff08\u5982 GPT-J\u3001GPT-NEO \u7b49\uff09\u65b9\u9762\u53d1\u6325\u4e86\u81f3\u5173\u91cd\u8981\u7684\u4f5c\u7528\u3002\u8fd9\u4e9b\u57fa\u7840\u6027\u5de5\u4f5c\u4e5f\u4e3a\u6570\u636e\u8fc7\u6ee4\u65b9\u6cd5\u548c\u7f29\u653e\u7684\u540e\u7eed\u6539\u8fdb\u94fa\u5e73\u4e86\u9053\u8def\u3002\u540c\u6837\uff0c\u5728\u56fe\u50cf\u6587\u672c\u9886\u57df\uff0c\u5927\u89c4\u6a21\u7684\u5f00\u6e90\u6570\u636e\u96c6\u523a\u6fc0\u4e86\u66f4\u597d\u7684\u6570\u636e\u7ba1\u7406\u65b9\u6cd5\u7684\u521b\u65b0\uff0c\u4f8b\u5982\u6570\u636e\u8fc7\u6ee4\u7f51\u7edc\u548cT-MARS\u3002\u4ece\u524d\u6cbf\u5b9e\u9a8c\u5ba4\u5230\u8bad\u7ec3\u5927\u578b\u591a\u6a21\u6001\u6a21\u578b \uff08LMM\uff09<\/a>\u00a0\u6709\u4e00\u4e2a\u660e\u663e\u7684\u8f6c\u53d8\uff0c\u8fd9\u9700\u8981\u5927\u91cf\u7684\u591a\u6a21\u6001\u4ea4\u9519\u6570\u636e\u96c6\uff0c\u5305\u62ec\u56fe\u50cf\u548c\u6587\u672c\u7684\u81ea\u7531\u683c\u5f0f\u5e8f\u5217\u3002\u968f\u7740\u524d\u6cbf\u6a21\u578b\u80fd\u529b\u7684\u5feb\u901f\u63d0\u5347\uff0c\u95ed\u6e90\u6a21\u578b\u548c\u5f00\u6e90\u6a21\u578b\u4e4b\u95f4\u7684\u591a\u6a21\u6001\u8bad\u7ec3\u6570\u636e\u51fa\u73b0\u4e86\u660e\u663e\u7684\u5dee\u8ddd\u3002\u5f53\u524d\u7684\u5f00\u6e90\u591a\u6a21\u6001\u4ea4\u9519\u6570\u636e\u96c6\u6bd4\u7eaf\u6587\u672c\u6570\u636e\u96c6\u66f4\u5c0f\uff0c\u591a\u6837\u6027\u66f4\u5c11\uff0c\u4e3b\u8981\u6765\u81eaHTML\u6587\u6863\uff0c\u8fd9\u9650\u5236\u4e86\u6570\u636e\u7684\u5e7f\u5ea6\u548c\u591a\u6837\u6027\u3002\u8fd9\u79cd\u9650\u5236\u963b\u788d\u4e86\u5f3a\u5927\u7684\u5f00\u6e90 LMM \u7684\u5f00\u53d1\uff0c\u5e76\u5728\u5f00\u6e90\u548c\u95ed\u6e90\u6a21\u578b\u7684\u80fd\u529b\u4e4b\u95f4\u9020\u6210\u4e86\u5dee\u5f02\u3002<\/p>\n \u4e3a\u4e86\u5f25\u8865\u8fd9\u4e00\u5dee\u8ddd\uff0cMINT-1T \u88ab\u521b\u5efa\u4e3a\u8fc4\u4eca\u4e3a\u6b62\u6700\u5927\u3001\u6700\u591a\u6837\u5316\u7684\u5f00\u6e90\u591a\u6a21\u6001\u4ea4\u9519\u6570\u636e\u96c6\u3002MINT-1T \u5305\u542b\u603b\u5171 1 \u4e07\u4ebf\u4e2a\u6587\u672c\u4ee4\u724c\u548c 30 \u4ebf\u5f20\u56fe\u50cf\uff0c\u8fd9\u4e9b\u56fe\u50cf\u6765\u81ea HTML\u3001PDF \u548c ArXiv \u7b49\u4e0d\u540c\u6765\u6e90\u3002\u5728 MINT-1T \u4e4b\u524d\uff0c\u8be5\u9886\u57df\u6700\u5927\u7684\u5f00\u6e90\u6570\u636e\u96c6\u662f OBELICS\uff0c\u5176\u4e2d\u5305\u62ec 1150 \u4ebf\u4e2a\u6587\u672c\u6807\u8bb0\u548c 3.53 \u4ebf\u5f20\u56fe\u50cf\uff0c\u5168\u90e8\u6765\u81ea HTML\u3002<\/p>\n MINT-1T\u7684\u8d21\u732e\u5982\u4e0b\uff1a<\/p>\n MINT-1T \u7b56\u5212\u4e86\u4e00\u4e2a\u5927\u89c4\u6a21\u7684\u5f00\u6e90\u6570\u636e\u96c6\uff0c\u8be5\u6570\u636e\u96c6\u5229\u7528\u4e86\u66f4\u591a\u6837\u5316\u7684\u4ea4\u9519\u6587\u6863\u6765\u6e90\uff0c\u4f8b\u5982 PDF \u548c ArXiv \u8bba\u6587\u3002\u672c\u8282\u8be6\u7ec6\u4ecb\u7ecd\u4e86 MINT-1T \u91c7\u8d2d\u591a\u6a21\u6001\u6587\u6863\u3001\u8fc7\u6ee4\u4f4e\u8d28\u91cf\u5185\u5bb9\u3001\u5220\u9664\u91cd\u590d\u6570\u636e\u4ee5\u53ca\u5220\u9664\u5bf9\u5de5\u4f5c\u6216 NSFW \u4e0d\u5b89\u5168\u4ee5\u53ca\u4e0d\u9700\u8981\u7684\u6750\u6599\u7684\u65b9\u6cd5\u3002\u6700\u7ec8\u6570\u636e\u96c6\u5305\u62ec 9220 \u4ebf \uff08B\uff09 \u4e2a HTML \u4ee4\u724c\u3001106B \u4e2a PDF \u4ee4\u724c\u548c 9B \u4e2a ArXiv \u4ee4\u724c\u3002<\/p>\n MINT-1T \u9075\u5faa OBELICS \u7684\u65b9\u6cd5\uff0c\u901a\u8fc7\u89e3\u6790\u6bcf\u4e2a WARC \u6761\u76ee\u7684 DOM \u6811\uff0c\u4ece CommonCrawl WARC \u6587\u4ef6\u4e2d\u63d0\u53d6\u4ea4\u9519\u7684\u591a\u6a21\u6001\u6587\u6863\u3002\u867d\u7136 OBELICS \u4ec5\u5904\u7406 2020 \u5e74 2 \u6708\u81f3 2023 \u5e74 2 \u6708\u7684 CommonCrawl \u8f6c\u50a8\u6587\u6863\uff0c\u4f46 MINT-1T \u5df2\u6269\u5c55\u6587\u6863\u6c60\u4ee5\u5305\u62ec 2017 \u5e74 5 \u6708\u81f3 2024 \u5e74 4 \u6708\u7684 HTML \u6587\u6863\uff082018 \u5e74 10 \u6708\u81f3 2024 \u5e74 4 \u6708\u7684\u5b8c\u6574\u8f6c\u50a8\u548c\u65e9\u5e74\u7684\u90e8\u5206\u8f6c\u50a8\uff09\u3002\u4e0e OBELICS \u7c7b\u4f3c\uff0cMINT-1T \u4f1a\u8fc7\u6ee4\u6389\u4e0d\u5305\u542b\u4efb\u4f55\u56fe\u50cf\u3001\u8d85\u8fc7 30 \u5f20\u56fe\u50cf\u6216\u4efb\u4f55 URL \u5305\u542b\u4e0d\u9002\u5f53\u5b50\u5b57\u7b26\u4e32\uff08\u5982\u5fbd\u6807\u3001\u5934\u50cf\u3001\u8272\u60c5\u548c xxx\uff09\u7684\u56fe\u50cf\u3002<\/p>\n MINT-1T \u4ece 2023 \u5e74 2 \u6708\u5230 2024 \u5e74 4 \u6708\u7684 CommonCrawl WAT \u6587\u4ef6\u4e2d\u83b7\u53d6 PDF \u6587\u6863\u3002\u6700\u521d\uff0c\u6240\u6709 PDF \u94fe\u63a5\u90fd\u662f\u4ece\u8fd9\u4e9b\u8f6c\u50a8\u4e2d\u63d0\u53d6\u7684\u3002\u7136\u540e\uff0cMINT-1T \u5c1d\u8bd5\u4f7f\u7528 PyMuPDF \u4e0b\u8f7d\u548c\u8bfb\u53d6 PDF\uff0c\u4e22\u5f03\u8d85\u8fc7 50MB\uff08\u53ef\u80fd\u5305\u542b\u5927\u56fe\u50cf\uff09\u548c\u8d85\u8fc7 50 \u9875\u7684 PDF\u3002\u6ca1\u6709\u6587\u672c\u7684\u9875\u9762\u5c06\u88ab\u6392\u9664\uff0c\u5e76\u4e3a\u5176\u4f59\u9875\u9762\u5efa\u7acb\u9605\u8bfb\u987a\u5e8f\u3002\u9605\u8bfb\u987a\u5e8f\u662f\u901a\u8fc7\u627e\u5230\u9875\u9762\u4e0a\u6240\u6709\u6587\u672c\u5757\u7684\u8fb9\u754c\u6846\uff0c\u6839\u636e\u5217\u5bf9\u5757\u8fdb\u884c\u805a\u7c7b\uff0c\u7136\u540e\u4ece\u5de6\u4e0a\u89d2\u5230\u53f3\u4e0b\u89d2\u5bf9\u5b83\u4eec\u8fdb\u884c\u6392\u5e8f\u6765\u786e\u5b9a\u7684\u3002\u56fe\u50cf\u6839\u636e\u5b83\u4eec\u4e0e\u540c\u4e00\u9875\u9762\u4e0a\u7684\u6587\u672c\u5757\u7684\u63a5\u8fd1\u7a0b\u5ea6\u96c6\u6210\u5230\u5e8f\u5217\u4e2d\u3002<\/p>\n MINT-1T \u4f7f\u7528 TexSoup \u4ece LaTeX \u6e90\u4ee3\u7801\u6784\u5efa ArXiv \u4ea4\u9519\u6587\u6863\uff0c\u4ee5\u67e5\u627e\u56fe\u5f62\u6807\u7b7e\u5e76\u5c06\u56fe\u50cf\u4e0e\u7eb8\u8d28\u6587\u672c\u4ea4\u9519\u3002\u5bf9\u4e8e\u591a\u6587\u4ef6\u8bba\u6587\uff0cMINT-1T \u4f1a\u8bc6\u522b\u4e3b Tex \u6587\u4ef6\uff0c\u5e76\u7528\u5176\u6587\u4ef6\u7684\u5185\u5bb9\u66ff\u6362\u8f93\u5165\u6807\u7b7e\u3002LaTeX \u4ee3\u7801\u901a\u8fc7\u5220\u9664\u5bfc\u5165\u3001\u53c2\u8003\u4e66\u76ee\u3001\u8868\u683c\u548c\u5f15\u6587\u6807\u7b7e\u8fdb\u884c\u6e05\u7406\u3002\u7531\u4e8e ArXiv \u5df2\u7ecf\u662f\u4e00\u4e2a\u9ad8\u5ea6\u7b56\u5212\u7684\u6570\u636e\u6e90\uff0c\u56e0\u6b64\u65e0\u9700\u6267\u884c\u989d\u5916\u7684\u8fc7\u6ee4\u548c\u91cd\u590d\u6570\u636e\u5220\u9664\u3002<\/p>\n MINT-1T \u9075\u5faa RefinedWeb\u3001Dolma \u548c FineWeb \u5efa\u7acb\u7684\u505a\u6cd5\uff0c\u907f\u514d\u4f7f\u7528\u57fa\u4e8e\u6a21\u578b\u7684\u542f\u53d1\u5f0f\u65b9\u6cd5\u8fdb\u884c\u6587\u672c\u8fc7\u6ee4\u3002\u6700\u521d\uff0c\u4f7f\u7528 Fasttext \u7684\u8bed\u8a00\u8bc6\u522b\u6a21\u578b\uff08\u7f6e\u4fe1\u5ea6\u9608\u503c\u4e3a 0.65\uff09\u6d88\u9664\u975e\u82f1\u8bed\u6587\u6863\u3002\u5982\u679c URL \u5305\u542b NSFW \u5b50\u5b57\u7b26\u4e32\uff0c\u5219\u8fd8\u4f1a\u5220\u9664\u6587\u6863\uff0c\u4ee5\u6392\u9664\u8272\u60c5\u5185\u5bb9\u548c\u4e0d\u826f\u5185\u5bb9\u3002\u5e94\u7528 RefinedWeb \u4e2d\u7684\u6587\u672c\u8fc7\u6ee4\u65b9\u6cd5\uff0c\u4e13\u95e8\u5220\u9664\u5177\u6709\u8fc7\u591a\u91cd\u590d n-gram \u7684\u6587\u6863\u6216\u4f7f\u7528 MassiveText \u89c4\u5219\u8bc6\u522b\u4e3a\u4f4e\u8d28\u91cf\u7684\u6587\u6863\u3002<\/p>\n \u5728\u6574\u7406 PDF \u548c HTML \u6587\u4ef6\u540e\uff0cMINT-1T \u5c1d\u8bd5\u4e0b\u8f7d HTML \u6570\u636e\u96c6\u4e2d\u7684\u6240\u6709\u56fe\u50cf URL\uff0c\u4e22\u5f03\u4e0d\u53ef\u68c0\u7d22\u7684\u94fe\u63a5\u5e76\u5220\u9664\u6ca1\u6709\u6709\u6548\u56fe\u50cf\u94fe\u63a5\u7684\u6587\u6863\u3002\u5c0f\u4e8e 150 \u50cf\u7d20\u7684\u56fe\u50cf\u5c06\u88ab\u4e22\u5f03\u4ee5\u907f\u514d\u5fbd\u6807\u548c\u56fe\u6807\u7b49\u5608\u6742\u7684\u56fe\u50cf\uff0c\u5927\u4e8e 20,000 \u50cf\u7d20\u7684\u56fe\u50cf\u4e5f\u4f1a\u88ab\u5220\u9664\uff0c\u56e0\u4e3a\u5b83\u4eec\u901a\u5e38\u5bf9\u5e94\u4e8e\u504f\u79bb\u4e3b\u9898\u7684\u56fe\u50cf\u3002\u5bf9\u4e8e HTML \u6587\u6863\uff0c\u5c06\u5220\u9664\u5bbd\u9ad8\u6bd4\u5927\u4e8e 2 \u7684\u56fe\u50cf\uff0c\u4ee5\u8fc7\u6ee4\u6389\u4f4e\u8d28\u91cf\u7684\u56fe\u50cf\uff0c\u4f8b\u5982\u5e7f\u544a\u6a2a\u5e45\u3002\u5bf9\u4e8e PDF\uff0c\u9608\u503c\u8c03\u6574\u4e3a 3 \u4e2a\uff0c\u4ee5\u4fdd\u7559\u79d1\u5b66\u7684\u6570\u5b57\u548c\u8868\u683c\u3002<\/p>\n MINT-1T \u5728\u6bcf\u4e2a CommonCrawl \u5feb\u7167\u548c\u56fe\u50cf\u91cd\u590d\u6570\u636e\u5220\u9664\u4e2d\u6267\u884c\u6bb5\u843d\u548c\u6587\u6863\u6587\u672c\u91cd\u590d\u6570\u636e\u5220\u9664\uff0c\u4ee5\u5220\u9664\u91cd\u590d\u7684\u3001\u65e0\u4fe1\u606f\u7684\u56fe\u50cf\uff0c\u4f8b\u5982\u56fe\u6807\u548c\u5fbd\u6807\u3002\u6240\u6709\u91cd\u590d\u6570\u636e\u5220\u9664\u6b65\u9aa4\u90fd\u9488\u5bf9\u6bcf\u4e2a\u6570\u636e\u6e90\u5355\u72ec\u6267\u884c\u3002<\/p>\n \u9075\u5faa Dolma \u7684\u65b9\u6cd5\uff0cMINT-1T \u4f7f\u7528 Bloom Filter \u8fdb\u884c\u9ad8\u6548\u7684\u6587\u672c\u91cd\u590d\u6570\u636e\u5220\u9664\uff0c\u5c06\u8bef\u62a5\u7387\u8bbe\u7f6e\u4e3a 0.01\uff0c\u5e76\u5bf9\u6bcf\u4e2a\u6587\u6863\u7684 13 \u514b\u6bb5\u843d\uff08\u901a\u8fc7\u53cc\u6362\u884c\u7b26\u5206\u9694\u7b26\u8868\u793a\uff09\u8fdb\u884c\u91cd\u590d\u6570\u636e\u5220\u9664\u3002\u5982\u679c\u6587\u6863\u4e2d 80% \u4ee5\u4e0a\u7684\u6bb5\u843d\u662f\u91cd\u590d\u7684\uff0c\u5219\u6574\u4e2a\u6587\u6863\u5c06\u88ab\u4e22\u5f03\u3002<\/p>\n \u6bb5\u843d\u91cd\u590d\u6570\u636e\u5220\u9664\u540e\uff0cMINT-1T \u4f1a\u5220\u9664 HTML \u6587\u6863\u4e2d\u5e38\u89c1\u7684\u7b80\u77ed\u6837\u677f\u53e5\u5b50\uff0c\u4f8b\u5982\u201c\u8df3\u5230\u5185\u5bb9\u201d\u6216\u201c\u535a\u5ba2\u5b58\u6863\u201d\u3002\u8fd9\u662f\u901a\u8fc7\u5bf9\u6bcf\u4e2a CommonCrawl \u5feb\u7167\u7684 2% \u8fd0\u884c\u7cbe\u786e\u7684\u6bb5\u843d\u91cd\u590d\u6570\u636e\u5220\u9664\u6765\u5b9e\u73b0\u7684\uff0c\u7b26\u5408 CCNet \u7684\u505a\u6cd5\uff0c\u786e\u4fdd\u5927\u90e8\u5206\u5185\u5bb9\u5220\u9664\u5e38\u89c1\u7684\u6837\u677f\u6587\u672c\u3002<\/p>\n \u5728\u6bcf\u4e2a CommonCrawl \u5feb\u7167\u4e2d\uff0cMINT-1T \u4f1a\u6839\u636e SHA256 \u54c8\u5e0c\u5220\u9664\u9891\u7e41\u51fa\u73b0\u7684\u56fe\u50cf\u3002\u6309\u7167\u591a\u6a21\u5f0f C4 \u505a\u6cd5\uff0c\u4ec5\u5220\u9664\u5728\u5feb\u7167\u4e2d\u51fa\u73b0\u8d85\u8fc7 10 \u6b21\u7684\u56fe\u50cf\uff0c\u800c\u4e0d\u662f\u4e25\u683c\u7684\u91cd\u590d\u6570\u636e\u5220\u9664\u3002\u4e0e OBELICS \u4e00\u81f4\uff0c\u5355\u4e2a\u6587\u6863\u4e2d\u7684\u91cd\u590d\u56fe\u50cf\u5c06\u88ab\u5220\u9664\uff0c\u4ec5\u4fdd\u7559\u7b2c\u4e00\u6b21\u51fa\u73b0\u3002<\/p>\n \u5728\u6574\u4e2a\u6570\u636e\u5904\u7406\u8fc7\u7a0b\u4e2d\uff0cMINT-1T \u5e73\u5747\u53ef\u4ee5\u4ece 190 \u5904\u7406\u5668\u548c 90 \u5904\u7406\u5668\u8282\u70b9\u7684\u6df7\u5408\u4e2d\u8bbf\u95ee 2,350 \u4e2a CPU \u5185\u6838\u3002\u603b\u5171\u4f7f\u7528\u4e86\u5927\u7ea6 420 \u4e07\u4e2a CPU \u5c0f\u65f6\u6765\u6784\u5efa\u6b64\u6570\u636e\u96c6\u3002<\/p>\n\n\n
\n <\/td>\n<\/tr>\n \n <\/td>\n <\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n \n
MINT-1T\uff1a\u6784\u5efa\u6570\u636e\u96c6<\/h2>\n
\u91c7\u8d2d\u5927\u91cf\u591a\u5f0f\u8054\u8fd0\u6587\u4ef6<\/h3>\n
HTML \u7ba1\u9053<\/h3>\n
PDF \u7ba1\u7ebf<\/h3>\n
\n\n
\n <\/td>\n<\/tr>\n \n <\/td>\n <\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n ArXiv \u6d41\u6c34\u7ebf<\/h3>\n
\u6587\u672c\u8d28\u91cf\u8fc7\u6ee4<\/h3>\n
\u56fe\u50cf\u8fc7\u6ee4<\/h3>\n
\n\n
\n <\/td>\n<\/tr>\n \n <\/td>\n <\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n \u5b89\u5168\u8fc7\u6ee4<\/h3>\n
\n
\u91cd\u590d\u6570\u636e\u5220\u9664<\/h3>\n
\u6bb5\u843d\u548c\u6587\u6863\u91cd\u590d\u6570\u636e\u5220\u9664<\/h4>\n
\u5220\u9664\u5e38\u89c1\u7684\u6837\u677f\u6587\u672c<\/h4>\n
\n\n
\n <\/td>\n<\/tr>\n \n <\/td>\n <\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n \u56fe\u50cf\u91cd\u590d\u6570\u636e\u5220\u9664<\/h4>\n
\u57fa\u7840\u8bbe\u65bd<\/h3>\n
\u6bd4\u8f83 MINT-1T \u4e0e OBELICS \u4e2d\u7684\u6587\u6863\u7ec4\u6210<\/h3>\n
\n\n
\n <\/td>\n<\/tr>\n \n <\/td>\n <\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n