<\/section>\n <\/p>\n
2024\u5e7408\u670823\u65e5 00:00<\/em>\u00a0<\/span>\u56db\u5ddd<\/span><\/em><\/span><\/p>\n<\/section>\n\n\n\n\n\n\n\u963f\u91cc\u5df4\u5df4\u96c6\u56e2\u7684\u7814\u7a76\u4eba\u5458\u5f00\u53d1\u4e86\u591a\u6a21\u6001\u8bed\u97f3\u4ea4\u4e92\u6846\u67b6FunAudioLLM\uff0c\u8be5\u6846\u67b6\u7531\u8bed\u97f3\u8bc6\u522b\u548c\u60c5\u611f\u5206\u6790\u6a21\u578bSenseVoice\u548c\u81ea\u7136\u8bed\u97f3\u751f\u6210\u6a21\u578bCosyVoice\u7ec4\u6210\uff0c\u5b9e\u73b0\u4e86\u4f4e\u5ef6\u8fdf\u3001\u9ad8\u7cbe\u5ea6\u3001\u591a\u8bed\u8a00\u548c\u60c5\u611f\u4e30\u5bcc\u7684\u8bed\u97f3\u4ea4\u4e92\u4f53\u9a8c\uff0c\u5e76\u5728\u8bed\u97f3\u7ffb\u8bd1\u3001\u60c5\u611f\u8bed\u97f3\u804a\u5929\u3001\u4e92\u52a8\u64ad\u5ba2\u548c\u6709\u58f0\u4e66\u65c1\u767d\u7b49\u9886\u57df\u5c55\u73b0\u51fa\u5e7f\u6cdb\u7684\u5e94\u7528\u524d\u666f\u3002<\/span><\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n\n\u8bba\u6587\u4ecb\u7ecd<\/strong><\/span><\/h2>\n<\/section>\n\u8bed\u97f3\u4ea4\u4e92\u6280\u672f\u968f\u7740\u4eba\u5de5\u667a\u80fd (AI) \u7684\u8fdb\u6b65\u800c\u663e\u8457\u53d1\u5c55\u3002\u8be5\u9886\u57df\u4e13\u6ce8\u4e8e\u589e\u5f3a\u4eba\u673a\u4e4b\u95f4\u7684\u81ea\u7136\u4ea4\u6d41\uff0c\u65e8\u5728\u4f7f\u4ea4\u4e92\u66f4\u52a0\u76f4\u89c2\u548c\u4eba\u6027\u5316\u3002\u6700\u8fd1\u7684\u53d1\u5c55\u4f7f\u5f97\u5b9e\u73b0\u9ad8\u7cbe\u5ea6\u8bed\u97f3\u8bc6\u522b\u3001\u60c5\u7eea\u68c0\u6d4b\u548c\u81ea\u7136\u8bed\u97f3\u751f\u6210\u6210\u4e3a\u53ef\u80fd\u3002\u7814\u7a76\u4eba\u5458\u4e00\u76f4\u5728\u521b\u5efa\u80fd\u591f\u5904\u7406\u591a\u79cd\u8bed\u8a00\u548c\u7406\u89e3\u60c5\u7eea\u7684\u6a21\u578b\uff0c\u4f7f\u4ea4\u4e92\u66f4\u52a0\u65e0\u7f1d\u548c\u4eba\u6027\u5316\u3002<\/section>\n\u8bed\u97f3\u4ea4\u4e92\u7684\u4e3b\u8981\u6311\u6218\u5728\u4e8e\u589e\u5f3a\u4e0e\u5927\u578b\u8bed\u8a00\u6a21\u578b (LLM) \u7684\u81ea\u7136\u8bed\u97f3\u4ea4\u4e92\u3002\u5f53\u524d\u7684\u7cfb\u7edf\u5728\u5ef6\u8fdf\u3001\u591a\u8bed\u8a00\u652f\u6301\u4ee5\u53ca\u751f\u6210\u5177\u6709\u60c5\u611f\u5171\u9e23\u548c\u4e0a\u4e0b\u6587\u9002\u5f53\u7684\u8bed\u97f3\u65b9\u9762\u901a\u5e38\u9700\u8981\u5e2e\u52a9\u3002\u8fd9\u4e9b\u9650\u5236\u963b\u788d\u4e86\u65e0\u7f1d\u548c\u4eba\u6027\u5316\u7684\u4ea4\u4e92\u3002\u589e\u5f3a\u8fd9\u4e9b\u7cfb\u7edf\u7406\u89e3\u548c\u5f00\u53d1\u8de8\u4e0d\u540c\u8bed\u8a00\u548c\u60c5\u611f\u73af\u5883\u7684\u51c6\u786e\u8bed\u97f3\u7684\u80fd\u529b\u5bf9\u4e8e\u63a8\u8fdb\u4eba\u673a\u4ea4\u4e92\u81f3\u5173\u91cd\u8981\u3002<\/section>\n\u73b0\u6709\u7684\u8bed\u97f3\u4ea4\u4e92\u65b9\u6cd5\u5305\u62ec\u5404\u79cd\u8bed\u97f3\u8bc6\u522b\u548c\u751f\u6210\u6a21\u578b\u3002Whisper \u7b49\u7528\u4e8e\u8bed\u97f3\u8bc6\u522b\u7684\u5de5\u5177\u4ee5\u53ca\u7528\u4e8e\u60c5\u7eea\u68c0\u6d4b\u548c\u97f3\u9891\u4e8b\u4ef6\u5206\u7c7b\u7684\u4f20\u7edf\u6a21\u578b\u4e3a\u8bed\u97f3\u4ea4\u4e92\u5960\u5b9a\u4e86\u57fa\u7840\u3002\u7136\u800c\uff0c\u8fd9\u4e9b\u65b9\u6cd5\u901a\u5e38\u65e0\u6cd5\u63d0\u4f9b\u8de8\u591a\u79cd\u8bed\u8a00\u7684\u4f4e\u5ef6\u8fdf\u3001\u9ad8\u7cbe\u5ea6\u548c\u60c5\u611f\u4e30\u5bcc\u7684\u4ea4\u4e92\u3002\u663e\u7136\uff0c\u9700\u8981\u4e00\u4e2a\u66f4\u5f3a\u5927\u3001\u66f4\u901a\u7528\u7684\u89e3\u51b3\u65b9\u6848\u6765\u6709\u6548\u5730\u5904\u7406\u8fd9\u4e9b\u4efb\u52a1\u3002<\/section>\n![\"\"](\"https:\/\/linguaresources.com\/wp-content\/uploads\/2024\/08\/640-9-9.png\")
<\/section>\n\u963f\u91cc\u5df4\u5df4\u96c6\u56e2\u7684\u7814\u7a76\u4eba\u5458\u63a8\u51fa\u4e86 FunAudioLLM\uff0c\u5b83\u5305\u542b\u4e24\u4e2a\u6838\u5fc3\u6a21\u578b\uff1aSenseVoice \u548c CosyVoice\u3002SenseVoice \u5728\u591a\u8bed\u8a00\u8bed\u97f3\u8bc6\u522b\u3001\u60c5\u7eea\u8bc6\u522b\u548c\u97f3\u9891\u4e8b\u4ef6\u68c0\u6d4b\u65b9\u9762\u8868\u73b0\u51fa\u8272\uff0c\u652f\u6301 50 \u591a\u79cd\u8bed\u8a00\u3002CosyVoice \u4e13\u6ce8\u4e8e\u81ea\u7136\u8bed\u97f3\u751f\u6210\uff0c\u5141\u8bb8\u63a7\u5236\u8bed\u8a00\u3001\u97f3\u8272\u3001\u8bf4\u8bdd\u98ce\u683c\u548c\u8bf4\u8bdd\u4eba\u8eab\u4efd\u3002\u901a\u8fc7\u7ed3\u5408\u8fd9\u4e9b\u6a21\u578b\uff0c\u8be5\u7814\u7a76\u56e2\u961f\u65e8\u5728\u7a81\u7834\u8bed\u97f3\u4ea4\u4e92\u6280\u672f\u7684\u754c\u9650\u3002<\/section>\nFunAudioLLM \u80cc\u540e\u7684\u6280\u672f\u5efa\u7acb\u5728 SenseVoice \u548c CosyVoice \u7684\u5148\u8fdb\u67b6\u6784\u4e4b\u4e0a\u3002SenseVoice-Small \u4f7f\u7528\u975e\u81ea\u56de\u5f52\u6a21\u578b\u8fdb\u884c\u4e94\u79cd\u8bed\u8a00\u7684\u4f4e\u5ef6\u8fdf\u8bed\u97f3\u8bc6\u522b\uff0c\u5176\u6027\u80fd\u6bd4 Whisper-small \u5feb\u4e94\u500d\u4ee5\u4e0a\uff0c\u6bd4 Whisper-large \u5feb\u5341\u4e94\u500d\u4ee5\u4e0a\u3002SenseVoice-Large \u652f\u6301 50 \u591a\u79cd\u8bed\u8a00\u7684\u8bed\u97f3\u8bc6\u522b\uff0c\u63d0\u4f9b\u9ad8\u7cbe\u5ea6\u5e76\u652f\u6301\u60c5\u7eea\u8bc6\u522b\u548c\u97f3\u9891\u4e8b\u4ef6\u68c0\u6d4b\u7b49\u590d\u6742\u4efb\u52a1\u3002CosyVoice \u91c7\u7528\u76d1\u7763\u8bed\u4e49\u8bed\u97f3\u6807\u8bb0\u8fdb\u884c\u81ea\u7136\u4e14\u5177\u6709\u60c5\u611f\u8868\u8fbe\u529b\u7684\u8bed\u97f3\u751f\u6210\uff0c\u80fd\u591f\u8fdb\u884c\u96f6\u6837\u672c\u5b66\u4e60\u548c\u8de8\u8bed\u8a00\u8bed\u97f3\u514b\u9686\u3002<\/section>\n![\"\"](\"https:\/\/linguaresources.com\/wp-content\/uploads\/2024\/08\/640-10-10.png\")
<\/section>\nFunAudioLLM \u7684\u6027\u80fd\u4e0e\u73b0\u6709\u6a21\u578b\u76f8\u6bd4\u6709\u663e\u7740\u6539\u8fdb\u3002SenseVoice \u6bd4 Whisper \u5b9e\u73b0\u4e86\u66f4\u5feb\u3001\u66f4\u51c6\u786e\u7684\u8bed\u97f3\u8bc6\u522b\u3002\u4f8b\u5982\uff0cSenseVoice-Small \u7684\u8bc6\u522b\u5ef6\u8fdf\u5c0f\u4e8e 80 \u6beb\u79d2\uff0c\u8fdc\u4f4e\u4e8e\u5176\u5bf9\u5e94\u6a21\u578b\u3002\u4e0e Whisper \u76f8\u6bd4\uff0cSenseVoice-Large \u5728\u591a\u79cd\u8bed\u8a00\u4e2d\u8868\u73b0\u51fa\u9ad8\u7cbe\u5ea6\u7684\u81ea\u52a8\u8bed\u97f3\u8bc6\u522b (ASR)\uff0c\u8bcd\u9519\u8bef\u7387 (WER) \u964d\u4f4e\u4e86 20% \u4ee5\u4e0a\u3002CosyVoice \u64c5\u957f\u751f\u6210\u9488\u5bf9\u7279\u5b9a\u8bf4\u8bdd\u4eba\u7684\u591a\u8bed\u8a00\u8bed\u97f3\uff0cWER \u4f4e\u4e8e 2%\uff0c\u8bf4\u8bdd\u4eba\u76f8\u4f3c\u5ea6\u5f97\u5206\u8d85\u8fc7 75%\uff0c\u4e0e\u4eba\u7c7b\u76f8\u5f53\u3002\u5b83\u652f\u6301\u96f6\u6837\u672c\u4e0a\u4e0b\u6587\u5b66\u4e60\uff0c\u53ea\u9700\u4e09\u79d2\u7684\u63d0\u793a\u5373\u53ef\u8fdb\u884c\u8bed\u97f3\u514b\u9686\uff0c\u5e76\u901a\u8fc7\u6307\u4ee4\u6587\u672c\u63d0\u4f9b\u5bf9\u8bed\u97f3\u8f93\u51fa\u7684\u8be6\u7ec6\u63a7\u5236\u3002<\/section>\n![\"\"](\"https:\/\/linguaresources.com\/wp-content\/uploads\/2024\/08\/640-11-11.png\")
<\/section>\n\u6700\u540e\uff0c\u963f\u91cc\u5df4\u5df4\u96c6\u56e2\u7684\u7814\u7a76\u4eba\u5458\u5c55\u793a\u4e86 FunAudioLLM \u53ef\u4ee5\u901a\u8fc7\u5404\u79cd\u5b9e\u7528\u65b9\u5f0f\u5e94\u7528\u3002\u8fd9\u4e9b\u5e94\u7528\u5305\u62ec\u8bed\u97f3\u5230\u8bed\u97f3\u7ffb\u8bd1\uff0c\u4f7f\u7528\u6237\u80fd\u591f\u4f7f\u7528\u81ea\u5df1\u7684\u58f0\u97f3\u8bf4\u5916\u8bed\uff1b\u60c5\u611f\u8bed\u97f3\u804a\u5929\uff0c\u6a21\u578b\u53ef\u4ee5\u7406\u89e3\u548c\u56de\u5e94\u60c5\u7eea\uff0c\u5b9e\u73b0\u66f4\u4eba\u6027\u5316\u7684\u4e92\u52a8\uff1b\u4e92\u52a8\u64ad\u5ba2\uff0c\u5141\u8bb8\u7528\u6237\u4e0e\u591a\u4e2a\u5927\u578b\u6a21\u578b\u8fdb\u884c\u5b9e\u65f6\u8ba8\u8bba\uff1b\u4ee5\u53ca\u5bcc\u6709\u8868\u73b0\u529b\u7684\u6709\u58f0\u8bfb\u7269\u65c1\u767d\uff0c\u4e3a\u6709\u58f0\u8bfb\u7269\u63d0\u4f9b\u591a\u89d2\u8272\u65c1\u767d\u3002SenseVoice \u548c CosyVoice \u4e0e LLM \u7684\u96c6\u6210\u5b9e\u73b0\u4e86\u8fd9\u4e9b\u9ad8\u7ea7\u529f\u80fd\uff0c\u5c55\u793a\u4e86 FunAudioLLM \u5728\u7a81\u7834\u8bed\u97f3\u4ea4\u4e92\u6280\u672f\u8fb9\u754c\u65b9\u9762\u7684\u6f5c\u529b\u3002<\/section>\n\n\u8bba\u6587\u4e0b\u8f7d<\/strong><\/span><\/h2>\n<\/section>\n\n- \n
\u8bba\u6587<\/strong>\u5730\u5740\uff1ahttps:\/\/arxiv.org\/abs\/2407.04051<\/p>\n<\/li>\n- \n
Github<\/strong>\u5730\u5740\uff1ahttps:\/\/github.com\/FunAudioLLM<\/p>\n<\/li>\n<\/ul>\n<\/section>\n\n\n<\/section>\n\u58f0\u660e\uff1a\u672c\u516c\u4f17\u53f7\u8f6c\u8f7d\u6b64\u6587\u7ae0\u662f\u51fa\u4e8e\u4f20\u64ad\u884c\u4e1a\u8d44\u8baf\u3001\u6d1e\u89c1\u4e4b\u76ee\u7684\uff0c\u5982\u6709\u4fb5\u72af\u5230\u60a8\u7684\u5408\u6cd5\u6743\u76ca\uff0c\u8bf7\u81f4\u4fe1\uff1a532541801@qq.com\uff0c\u6211\u4eec\u5c06\u53ca\u65f6\u8c03\u6574\u5904\u7406\u3002\u8c22\u8c22\u652f\u6301\uff01<\/span><\/section>\n\n\n\n\n\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n\n\n\n![\"\"](\"https:\/\/linguaresources.com\/wp-content\/uploads\/2024\/08\/640-12-8-300x300.png\")
<\/section>\n<\/section>\n\n\n\n\n\n\u3010\u8bed\u8a00\u670d\u52a1\u884c\u4e1a\u3011\u5206\u4eab\u7fa4<\/strong><\/section>\n<\/section>\n<\/section>\n<\/section>\n\u7fa4\u5185\u4f1a\u5b9a\u671f\u63a8\u9001\u8bed\u8a00\u670d\u52a1\u884c\u4e1a\u6700\u65b0\u52a8\u6001\u3001\u6d3b\u52a8\u9884\u544a\u3001\u7ade\u8d5b\u901a\u77e5\ud83d\udcdd\u7b49\u5185\u5bb9~<\/span><\/span><\/p>\n\u00a0<\/span><\/span><\/p>\n\u6b22\u8fce\u4f60\u7684\u52a0\u5165\ud83e\udd70\uff01<\/span><\/span><\/p>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n\n![\"\"](\"https:\/\/linguaresources.com\/wp-content\/uploads\/2024\/08\/640-13-9.png\")
<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n<\/section>\n\u00a0<\/span><\/strong><\/section>\n-END-<\/span><\/strong><\/section>\n<\/section>\n","protected":false},"excerpt":{"rendered":"\u6280\u672f\u5e94\u7528\uff5cFunAudioLLM\uff1a\u6253\u7834\u8bed\u8a00\u58c1\u5792\uff0c\u5f00\u542f\u591a\u8bed\u97f3\u4ea4\u6d41\uff01 \u8bed\u8a00\u670d\u52a1\u884c\u4e1a 2024\u5e7408\u6708 […]<\/p>\n","protected":false},"author":1,"featured_media":32709,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[82],"tags":[],"class_list":["post-32708","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-82"],"_links":{"self":[{"href":"https:\/\/linguaresources.com\/index.php?rest_route=\/wp\/v2\/posts\/32708","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/linguaresources.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/linguaresources.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/linguaresources.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/linguaresources.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=32708"}],"version-history":[{"count":2,"href":"https:\/\/linguaresources.com\/index.php?rest_route=\/wp\/v2\/posts\/32708\/revisions"}],"predecessor-version":[{"id":32715,"href":"https:\/\/linguaresources.com\/index.php?rest_route=\/wp\/v2\/posts\/32708\/revisions\/32715"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/linguaresources.com\/index.php?rest_route=\/wp\/v2\/media\/32709"}],"wp:attachment":[{"href":"https:\/\/linguaresources.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=32708"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/linguaresources.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=32708"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/linguaresources.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=32708"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}