{"id":10006641,"date":"2022-09-30T21:52:13","date_gmt":"2022-09-30T13:52:13","guid":{"rendered":"https:\/\/ityet.com\/?p=10006641"},"modified":"2022-09-30T21:52:13","modified_gmt":"2022-09-30T13:52:13","slug":"easynlp%e7%8e%a9%e8%bd%ac%e6%96%87%e6%9c%ac%e6%91%98%e8%a6%81%ef%bc%88%e6%96%b0%e9%97%bb%e6%a0%87%e9%a2%98%ef%bc%89%e7%94%9f%e6%88%90","status":"publish","type":"post","link":"https:\/\/itmax.cn\/?p=10006641","title":{"rendered":"EasyNLP\u73a9\u8f6c\u6587\u672c\u6458\u8981\uff08\u65b0\u95fb\u6807\u9898\uff09\u751f\u6210"},"content":{"rendered":"<p>\u7b80\u4ecb\uff1a\u00a0\u672c\u2f42\u5c06\u63d0\u4f9b\u5173\u4e8ePEGASUS\u7684\u6280\u672f\u89e3\u8bfb\uff0c\u4ee5\u53ca\u5982\u4f55\u5728EasyNLP\u6846\u67b6\u4e2d\u4f7f\u2f64\u4e0ePEGASUS\u76f8\u5173\u7684\u6587\u672c\u6458\u8981\uff08\u65b0\u95fb\u6807\u9898\uff09\u751f\u6210\u6a21\u578b\u3002\u4f5c\u8005\uff1a\u738b\u660e\u3001\u9ec4\u4fca\u00a0\u5bfc\u8bfb\u6587\u672c\u751f\u6210\u662f\u81ea\u7136\u8bed\u8a00\u5904\u7406\u9886\u57df\u7684\u4e00\u4e2a\u91cd\u8981\u7814\u7a76\u65b9\u5411\uff0c\u5177\u6709\u4e30\u5bcc\u7684\u5b9e\u9645\u5e94\u7528\u573a\u666f\u4ee5\u53ca\u7814\u7a76\u4ef7\u503c\u3002\u5176\u4e2d\uff0c\u751f\u6210\u5f0f\u6587\u672c\u6458\u8981\u4f5c\u4e3a\u6587\u672c\u751f\u6210\u7684\u4e00\u4e2a\u91cd\u8981\u5b50\u4efb\u52a1\uff0c\u5728\u5b9e\u9645\u5e94\u7528\u573a\u666f\u4e2d\uff0c\u5305\u62ec\u65b0\u95fb\u6807\u9898\u751f\u6210\u3001\u6458\u8981\u751f\u6210\u3001\u5173\u952e\u8bcd\u751f\u6210\u7b49\u4efb\u52a1\u5f62\u5f0f\u3002\u9884\u8bad\u7ec3\u8bed\u8a00\u6a21\u578b\uff0c\u5982BERT\u3001MASS\u3001uniLM\u7b49\u867d\u7136\u5728NLU\u573a\u666f\u4e2d\u53d6\u5f97\u4e86\u4ee4\u4eba\u77a9\u76ee\u7684\u6027\u80fd\uff0c\u4f46\u6a21\u578b\u91c7\u7528\u7684\u5355\u8bcd\u3001\u5b50\u8bcd\u906e\u76d6\u8bed\u8a00\u6a21\u578b\u5e76\u4e0d\u9002\u7528\u4e8e\u6587\u672c\u751f\u6210\u573a\u666f\u4e2d\uff0c\u7279\u522b\u662f\u751f\u6210\u5f0f\u6587\u672c\u6458\u8981\u573a\u666f\u3002\u5176\u539f\u56e0\u662f\uff0c\u751f\u6210\u5f0f\u6587\u672c\u6458\u8981\u4efb\u52a1\u5f80\u5f80\u8981\u6c42\u6a21\u578b\u5177\u6709\u66f4\u7c97\u7c92\u5ea6\u7684\u8bed\u4e49\u7406\u89e3\uff0c\u5982\u53e5\u5b50\u3001\u6bb5\u843d\u8bed\u4e49\u7406\u89e3\uff0c\u4ee5\u6b64\u8fdb\u884c\u6458\u8981\u751f\u6210\u3002\u4e3a\u4e86\u89e3\u51b3\u4e0a\u8ff0\u95ee\u9898\uff0cPEGASUS\u6a21\u578b\uff08PEGASUS: Pre-training with Extracted Gap-sentences for Abstractive Summarization\uff09\u9488\u5bf9\u6587\u672c\u6458\u8981\u4efb\u52a1\u8bbe\u8ba1\u4e86\u65e0\u76d1\u7763\u9884\u8bad\u7ec3\u4efb\u52a1\uff08Gap Sentence Generation\uff0c\u7b80\u79f0GSG\uff09\uff0c\u5373\u968f\u673a\u906e\u76d6\u6587\u6863\u4e2d\u7684\u51e0\u4e2a\u5b8c\u6574\u53e5\u5b50\uff0c\u8ba9\u6a21\u578b\u751f\u6210\u88ab\u906e\u76d6\u7684\u53e5\u5b50\u3002\u8be5\u9884\u8bad\u7ec3\u4efb\u52a1\u80fd\u591f\u5f88\u597d\u5730\u4e0e\u5b9e\u9645\u5730\u6587\u672c\u6458\u8981\u4efb\u52a1\u5339\u914d\uff0c\u4ece\u800c\u4f7f\u5f97\u9884\u8bad\u7ec3\u540e\u7684\u6a21\u578b\u7ecf\u8fc7\u7b80\u5355\u7684\u5fae\u8c03\u540e\u8fbe\u5230\u8f83\u597d\u7684\u6458\u8981\u751f\u6210\u6548\u679c\u3002\u56e0\u6b64\uff0c\u6211\u4eec\u5728EasyNLP\u6846\u67b6\u4e2d\u96c6\u6210\u4e86PEGASUS\u7b97\u6cd5\u548c\u6a21\u578b\uff0c\u4f7f\u7528\u6237\u80fd\u591f\u65b9\u4fbf\u5730\u4f7f\u7528\u8be5\u6a21\u578b\u8fdb\u884c\u6587\u672c\u6458\u8981\u751f\u6210\u76f8\u5173\u4efb\u52a1\u7684\u8bad\u7ec3\u548c\u9884\u6d4b\u3002EasyNLP\uff08<a href=\"https:\/\/link.segmentfault.com\/?enc=%2BoHj4yx0Q4EQCORkivleog%3D%3D.Zzs9UV%2FDHrwuMu6iN5BoZpB%2B0KyuNx6anT8n69y4%2BrDG1ppxqmUWDktIdCUkpVIi\" rel=\"nofollow\">https:\/\/github.com\/alibaba\/Ea&#8230;<\/a>\uff09\u662f\u963f\u2fa5\u4e91\u673a\u5668\u5b66\u4e60PAI \u56e2\u961f\u57fa\u4e8e PyTorch \u5f00\u53d1\u7684\u6613\u2f64\u4e14\u4e30\u5bcc\u7684\u4e2d\u2f42NLP\u7b97\u6cd5\u6846\u67b6\uff0c\u2f40\u6301\u5e38\u2f64\u7684\u4e2d\u2f42\u9884\u8bad\u7ec3\u6a21\u578b\u548c\u2f24\u6a21\u578b\u843d\u5730\u6280\u672f\uff0c\u5e76\u4e14\u63d0\u4f9b\u4e86\u4ece\u8bad\u7ec3\u5230\u90e8\u7f72\u7684\u2f00\u7ad9\u5f0f NLP \u5f00\u53d1\u4f53\u9a8c\u3002EasyNLP \u63d0\u4f9b\u4e86\u7b80\u6d01\u7684\u63a5\u2f1d\u4f9b\u2f64\u6237\u5f00\u53d1 NLP \u6a21\u578b\uff0c\u5305\u62ecNLP\u5e94\u2f64 AppZoo \u548c\u9884\u8bad\u7ec3 ModelZoo\uff0c\u540c\u65f6\u63d0\u4f9b\u6280\u672f\u5e2e\u52a9\u2f64\u6237\u2fbc\u6548\u7684\u843d\u5730\u8d85\u2f24\u9884\u8bad\u7ec3\u6a21\u578b\u5230\u4e1a\u52a1\u3002\u6587\u672c\u751f\u6210\u4f5c\u4e3a\u81ea\u7136\u8bed\u8a00\u5904\u7406\u7684\u4e00\u5927\u5b50\u4efb\u52a1\uff0c\u5177\u6709\u4f17\u591a\u7684\u5b9e\u9645\u5e94\u7528\uff0c\u5305\u62ec\u6807\u9898\u751f\u6210\u3001\u6587\u672c\u6458\u8981\u3001\u673a\u5668\u7ffb\u8bd1\u3001\u95ee\u7b54\u7cfb\u7edf\u3001\u5bf9\u8bdd\u7cfb\u7edf\u7b49\u7b49\u3002\u56e0\u6b64\uff0cEasyNLP\u4e5f\u5728\u9010\u6b65\u589e\u52a0\u5bf9\u6587\u672c\u751f\u6210\u5b50\u4efb\u52a1\u7684\u652f\u6301\uff0c\u5e0c\u671b\u80fd\u591f\u670d\u52a1\u66f4\u591a\u7684NLP\u4ee5\u53caNLG\u7b97\u6cd5\u5f00\u53d1\u8005\u548c\u7814\u7a76\u8005\uff0c\u4e5f\u5e0c\u671b\u548c\u793e\u533a\u4e00\u8d77\u63a8\u52a8NLG\u6280\u672f\u7684\u53d1\u5c55\u548c\u843d\u5730\u3002\u00a0\u672c\u2f42\u5c06\u63d0\u4f9b\u5173\u4e8ePEGASUS\u7684\u6280\u672f\u89e3\u8bfb\uff0c\u4ee5\u53ca\u5982\u4f55\u5728EasyNLP\u6846\u67b6\u4e2d\u4f7f\u2f64\u4e0ePEGASUS\u76f8\u5173\u7684\u6587\u672c\u6458\u8981\uff08\u65b0\u95fb\u6807\u9898\uff09\u751f\u6210\u6a21\u578b\u3002\u00a0Pegasus\u6a21\u578b\u8be6\u89e3<br \/><img decoding=\"async\" referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000042566823\" alt=\"\u56fe\u7247\" title=\"\u56fe\u7247\" loading=\"lazy\"><br \/>\u5728\u6b64\u4e4b\u524d\uff0c\u6587\u672c\u751f\u6210\u9884\u8bad\u7ec3\u6a21\u578bT5\u3001BART\u7b49\u6a21\u578b\u867d\u7136\u5728\u4f17\u591a\u6587\u672c\u751f\u6210\u4efb\u52a1\u4e2d\u53d6\u5f97\u4e86\u660e\u663e\u7684\u6027\u80fd\u589e\u76ca\uff0c\u4f46\u662f\u5728\u6587\u672c\u6458\u8981\u4efb\u52a1\u4e2d\uff0c\u6a21\u578b\u7684\u9884\u8bad\u7ec3\u76ee\u6807\u4e0e\u6587\u672c\u6458\u8981\u76ee\u6807\u8fd8\u662f\u5b58\u5728\u8f83\u5927\u7684\u5dee\u5f02\u3002\u8fd9\u5bfc\u81f4\u6b64\u7c7b\u9884\u8bad\u7ec3\u6a21\u578b\u5728\u8fc1\u79fb\u81f3\u4e0d\u7528\u9886\u57df\u7684\u6458\u8981\u4efb\u52a1\u65f6\uff0c\u4ecd\u7136\u9700\u8981\u8f83\u591a\u7684\u8bad\u7ec3\u6570\u636e\u5bf9\u6a21\u578b\u8fdb\u884c\u5fae\u8c03\u624d\u80fd\u8fbe\u5230\u8f83\u597d\u7684\u6548\u679c\u3002\u4e3a\u4e86\u7f13\u89e3\u4e0a\u8ff0\u95ee\u9898\uff0cPEGASUS\u6a21\u578b\u5728\u539f\u59cb\u7684\u5b50\u8bcd\u906e\u76d6\u635f\u5931\u7684\u57fa\u7840\u4e0a\uff0c\u589e\u52a0\u4e86\u5b8c\u6574\u53e5\u5b50\u906e\u76d6\u635f\u5931\uff0c\u5373\u5c06\u8f93\u5165\u6587\u6863\u4e2d\u7684\u968f\u673a\u51e0\u4e2a\u5b8c\u6574\u53e5\u5b50\u8fdb\u884c\u906e\u76d6\uff0c\u8ba9\u6a21\u578b\u590d\u539f\u3002\u5177\u4f53\u5730\uff0c\u5982\u4e0a\u56fe\u6240\u793a\uff0cPEGASUS\u91c7\u7528\u7f16\u7801\u5668-\u89e3\u7801\u5668\u67b6\u6784\uff08\u6807\u51c6transformer\u67b6\u6784\uff09\u3002\u6a21\u578b\u5bf9\u8f93\u5165\u91c7\u7528\u4e24\u79cd\u906e\u76d6\uff0c\u4e00\u79cd\u662fBERT\u91c7\u7528\u7684\u5b50\u8bcd\u906e\u76d6\uff0c\u7528\u3010mask2\u3011\u8868\u793a\uff0c\u8ba9\u6a21\u578b\u7684\u7f16\u7801\u5668\u8fd8\u539f\u88ab\u906e\u76d6\u7684\u5b50\u8bcd\uff08\u8be5\u7c7b\u635f\u5931\u5728\u6d88\u878d\u5b9e\u9a8c\u4e2d\u88ab\u8bc1\u660e\u5bf9\u4e0b\u6e38\u4efb\u52a1\u65e0\u6027\u80fd\u589e\u76ca\uff0c\u56e0\u6b64\u5728\u6700\u7ec8\u7684PEGASUS\u6a21\u578b\u4e2d\u5e76\u672a\u91c7\u7528\uff09\u3002\u53e6\u4e00\u79cd\u662fGSG\uff0c\u7528\u3010mask1\u3011\u8868\u793a\uff0c\u5373\u8ba9\u89e3\u7801\u5668\u751f\u6210\u8f93\u5165\u4e2d\u88ab\u906e\u76d6\u7684\u968f\u673a\u5b8c\u6574\u53e5\u5b50\u3002\u9488\u5bf9\u6b64\u635f\u5931\uff0c\u4f5c\u8005\u540c\u65f6\u63d0\u51fa\u4e09\u79cd\u53ef\u9009\u65b9\u6848\uff0c\u5305\u62ecRandom\uff08\u968f\u673a\u9009\u62e9m\u4e2a\u53e5\u5b50\uff09\u3001Lead\uff08\u9009\u62e9\u524dm\u4e2a\u53e5\u5b50\uff09\u3001Ind-Orig\uff08\u6839\u636e\u91cd\u8981\u6027\u5206\u6570\u9009\u62e9m\u4e2a\u53e5\u5b50\uff09\u3002\u5176\u4e2d\uff0c\u91cd\u8981\u6027\u5206\u6570\u5177\u4f53\u901a\u8fc7\u8ba1\u7b97\u6bcf\u53e5\u8bdd\u4e0e\u6587\u6863\u4e2d\u5176\u5b83\u53e5\u5b50\u96c6\u5408\u7684ROUGE\u5206\u6570\u5f97\u5230\u3002\u53ef\u4ee5\u8ba4\u4e3a\uff0c\u8be5\u7b56\u7565\u9009\u62e9\u80fd\u591f\u5f88\u5927\u7a0b\u5ea6\u4ee3\u8868\u6587\u6863\u4e2d\u5176\u5b83\u53e5\u5b50\u7684\u53e5\u5b50\u4f5c\u4e3a\u906e\u76d6\u5bf9\u8c61\u3002\u4e0b\u56fe\u5c55\u793a\u4e86\u4e09\u79cd\u9009\u53e5\u5b50\u65b9\u6848\u7684\u4e00\u4e2a\u4f8b\u5b50\uff0c\u6240\u9009\u53e5\u5b50\u5206\u522b\u88ab\u6807\u8bb0\u4e3a\u7eff\u8272\u3001\u7ea2\u68d5\u8272\u3001\u84dd\u8272\u3002\u5b9e\u9a8c\u8868\u660e\uff0c\u91c7\u7528\u7b2c\u4e09\u79cd\u53e5\u5b50\u9009\u62e9\u7b56\u7565\u7684\u6a21\u578b\u80fd\u591f\u53d6\u5f97\u6700\u4f18\u6027\u80fd\u3002<br \/><img decoding=\"async\" referrerpolicy=\"no-referrer\" src=\"https:\/\/segmentfault.com\/img\/remote\/1460000042566824\" alt=\"\u56fe\u7247\" title=\"\u56fe\u7247\" loading=\"lazy\"><br \/>\u00a0\u6587\u672c\u6458\u8981\u6a21\u578b\u4f7f\u7528\u6559\u7a0b\u4ee5\u4e0b\u6211\u4eec\u7b80\u8981\u4ecb\u7ecd\u5982\u4f55\u5728EasyNLP\u6846\u67b6\u4e2d\u4f7f\u7528PEGASUS\u4ee5\u53ca\u5176\u4ed6\u6587\u672c\u6458\u8981\u6a21\u578b\u3002\u00a0\u5b89\u88c5EasyNLP\u7528\u6237\u53ef\u4ee5\u76f4\u63a5\u53c2\u8003GitHub\uff08<a href=\"https:\/\/link.segmentfault.com\/?enc=3NTofd%2FE5PbvWBbvhXJl8A%3D%3D.Rqigrld6RtnWFhnUb9qRZYQUrRnlb7JUH1X1bY8W8a98bhto9tQqyePVkp9lrtcd\" rel=\"nofollow\">https:\/\/github.com\/alibaba\/Ea&#8230;<\/a>\uff09\u4e0a\u7684\u8bf4\u660e\u5b89\u88c5EasyNLP\u7b97\u6cd5\u6846\u67b6\u3002\u00a0\u6570\u636e\u51c6\u5907\u5728\u5177\u4f53\u7684\u6587\u672c\u6458\u8981\u573a\u666f\u4e2d\uff0c\u9700\u8981\u7528\u6237\u63d0\u4f9b\u4e0b\u6e38\u4efb\u52a1\u7684\u8bad\u7ec3\u4e0e\u9a8c\u8bc1\u6570\u636e\uff0c\u4e3atsv\u6587\u4ef6\u3002\u5bf9\u4e8e\u6587\u672c\u6458\u8981\u4efb\u52a1\uff0c\u8fd9\u4e00\u6587\u4ef6\u5305\u542b\u4ee5\u5236\u8868\u7b26\/t\u5206\u9694\u7684\u4e24\u5217\u6570\u636e\uff0c\u7b2c\u4e00\u5217\u662f\u6458\u8981\u5217\uff0c\u7b2c\u4e8c\u5217\u4e3a\u539f\u6587\u5217\u3002\u6837\u4f8b\u5982\u4e0b\uff1a\u6e56\u5317\uff1a\u201c\u56db\u4e0a\u4f01\u4e1a\u201d\u590d\u5de5\u7387\u5df2\u8fbe93.8% \u592e\u89c6\u7f51\u6d88\u606f\uff1a4\u67081\u65e5\uff0c\u8bb0\u8005\u4ece\u6e56\u5317\u7701\u65b0\u51a0\u80ba\u708e\u75ab\u60c5\u9632\u63a7\u5de5\u4f5c\u65b0\u95fb\u53d1\u5e03\u4f1a\u4e0a\u83b7\u6089\uff0c\u5728\u5404\u65b9\u9762\u5171\u540c\u52aa\u529b\u4e0b\uff0c\u6e56\u5317\u7701\u590d\u5de5\u590d\u4ea7\u5de5\u4f5c\u53d6\u5f97\u4e86\u9636\u6bb5\u6027\u6210\u6548\u3002\u622a\u81f33\u670831\u65e5\uff0c\u6e56\u5317\u7701\u201c\u56db\u4e0a\u4f01\u4e1a\u201d\u5305\u62ec\u89c4\u6a21\u4ee5\u4e0a\u5de5\u4e1a\u3001\u89c4\u6a21\u4ee5\u4e0a\u670d\u52a1\u4e1a\u6cd5\u4eba\u5355\u4f4d\u7b49\u7684\u590d\u5de5\u7387\u5df2\u8fbe93.8%\uff0c\u590d\u5c97\u738769.3%\u3002\u6b66\u6c49\u5e02\u7684\u590d\u5de5\u7387\u3001\u590d\u5c97\u7387\u4e5f\u5206\u522b\u8fbe\u5230\u4e8685.4%\u300140.4%\u3002\u8d23\u4efb\u7f16\u8f91\uff1a\u738b\u8bd7\u5c27\u00a0\u4e0b\u5217\u6587\u4ef6\u4e3a\u5df2\u7ecf\u5b8c\u6210\u9884\u5904\u7406\u7684\u65b0\u95fb\u6807\u9898\u751f\u6210\u8bad\u7ec3\u548c\u9a8c\u8bc1\u6570\u636e\uff0c\u53ef\u7528\u4e8e\u6d4b\u8bd5\uff1a<a href=\"https:\/\/link.segmentfault.com\/?enc=V%2BSXIUgS76VaTBmGwNRFDA%3D%3D.4zrjrqNXC2USU7g9E0qU%2F%2BM%2FREN%2FUDdWE5LTvHOq7nkHLadk3%2B4nhFZ7o08Ji%2F9YxHqTWRZTE49mrPWzy2z%2B43%2FSWILp4LWyWcCmvA7gpsCZmssVA1uf8y6dRJeBT34L\" rel=\"nofollow\">https:\/\/atp-modelzoo-sh.oss-c&#8230;<\/a>\u4e2d\u6587\u65b0\u95fb\u6807\u9898\u751f\u6210\u7531\u4e8ePEGASUS\u539f\u6587\u4ea7\u51fa\u7684\u6a21\u578b\u4ec5\u652f\u6301\u82f1\u6587\uff0c\u4e3a\u4e86\u65b9\u4fbf\u4e2d\u6587\u793e\u533a\u7528\u6237\u7684\u4f7f\u7528\uff0c\u6211\u4eec\u57fa\u4e8emT5\u7684\u6a21\u578b\u67b6\u6784\u9884\u8bad\u7ec3\u4e86\u4e00\u4e2a\u9488\u5bf9\u4e2d\u6587\u65b0\u95fb\u6807\u9898\u6458\u8981\u7684\u6a21\u578bmT5\uff0c\u5e76\u5c06\u5176\u96c6\u6210\u8fdbEasyNLP\u7684\u6a21\u578b\u5e93\u4e2d\u3002\u540c\u65f6\uff0c\u6211\u4eec\u8fd8\u96c6\u6210\u4e86IDEA\u673a\u6784\u9884\u8bad\u7ec3\u7684\u6587\u672c\u6458\u8981\u4e2d\u6587\u6a21\u578bRandeng\uff08\u53ef\u4ee5\u8ba4\u4e3a\u662f\u4e2d\u6587\u7248\u7684PEGASUS\uff09\uff0c\u4fbf\u4e8e\u7528\u6237\u63a2\u7d22\u4e0d\u540c\u6a21\u578b\u7684\u6027\u80fd\u3002\u4ee5\u4e0b\u6c47\u603b\u4e86EasyNLP\u4e2d\u53ef\u7528\u7684\u6a21\u578b\uff0c\u5e76\u5bf9\u6bd4\u6a21\u578b\u5728\u4e0a\u8ff0\u6570\u636e\u96c6\u4e0a\u7684\u6027\u80fd\u8868\u73b0\u3002\u63a8\u8350\u7528\u6237\u9009\u62e9\u524d\u4e24\u4e2a\u6a21\u578b\u8fdb\u884c\u6587\u672c\u6458\u8981\uff0c\u540e\u4e09\u4e2a\u6a21\u578b\u8fdb\u884c\u65b0\u95fb\u6807\u9898\u751f\u6210\u3002\u4e2d\u6587\u65b0\u95fb\u6807\u9898\uff08Rouge1\/2\/L\uff09\u8bba\u6587\u6807\u9898\u6458\u8981\uff08Rouge1\/2\/L\uff09hfl\/randeng-238M-Summary-Chinese59.66\/46.26\/55.9554.55\/39.37\/50.69hfl\/randeng-523M-Summary-Chinese62.86\/49.67\/58.8953.83\/39.17\/49.92alibaba-pai\/mt5-title-generation-zh-275m62.35\/48.63\/58.9654.28\/40.26\/50.55alibaba-pai\/randeng-238M-Summary-Chinese-tuned64.31\/51.80\/60.9758.83\/45.28\/55.72alibaba-pai\/randeng-523M-Summary-Chinese-tuned64.76\/51.65\/61.0659.27\/45.58\/55.92\u00a0\u5728\u65b0\u95fb\u6807\u9898\u751f\u6210\u4efb\u52a1\u4e2d\uff0c\u6211\u4eec\u91c7\u7528\u4ee5\u4e0b\u547d\u4ee4\u5bf9\u6a21\u578b\u8fdb\u884c\u8bad\u7ec3\u3002\u7528\u6237\u53ef\u4ee5\u6839\u636e\u8d85\u53c2\u6570\u2018save_checkpoint_steps\u2019\u6765\u51b3\u5b9a\u4fdd\u5b58\u6a21\u578b\u7684\u6b65\u6570\uff0c\u6846\u67b6\u5728\u6b64\u65f6\u4f1a\u5bf9\u8bad\u7ec3\u7684\u6a21\u578b\u8fdb\u884c\u8bc4\u6d4b\uff0c\u4f1a\u6839\u636e\u6a21\u578b\u5728\u9a8c\u8bc1\u96c6\u4e0a\u7684\u8868\u73b0\u51b3\u5b9a\u662f\u5426\u66f4\u65b0\u4fdd\u5b58\u7684\u6a21\u578b\u53c2\u6570\u3002\u5176\u4e2d\uff0c\u8fd0\u884c\u7684main.py\u6587\u4ef6\u5728EasyNLP\/examples\/appzoo_tutorials\/sequence_generation\u76ee\u5f55\u4e0b\uff0c\u540c\u65f6\u9700\u8981\u5c06\u8bad\u7ec3\u548c\u9a8c\u8bc1\u96c6\u6570\u636e\u653e\u5230\u8be5\u76ee\u5f55\u4e0b\u3002\u53ef\u4ee5\u5728\u2018user_defined_parameters\u2019\u8d85\u53c2\u6570\u4e0b\u7684\u2018pretrain_model_name_or_path\u2019\u6307\u5b9a\u4e0a\u8ff0\u8868\u683c\u4e2d\u7684\u6a21\u578b\u3002python main.py \/<\/p>\n<pre><code>--mode train \/\n--app_name=sequence_generation \/\n--worker_gpu=1 \/\n--tables=.\/cn_train.tsv,.\/cn_dev.tsv  \/\n--input_schema=title_tokens:str:1,content_tokens:str:1 \/\n--first_sequence=content_tokens \/\n--second_sequence=title_tokens \/\n--label_name=title_tokens \/\n--checkpoint_dir=.\/finetuned_zh_model \/\n--micro_batch_size=8 \/\n--sequence_length=512 \/\n--epoch_num=1  \/\n--save_checkpoint_steps=150 \/\n--export_tf_checkpoint_type none \/\n--user_defined_parameters 'pretrain_model_name_or_path=alibaba-pai\/mt5-title-generation-zh language=zh copy=false max_encoder_length=512 min_decoder_length=12 max_decoder_length=32 no_repeat_ngram_size=2 num_beams=5 num_return_sequences=5'<\/code><\/pre>\n<p>\u53e6\u5916\uff0c\u7528\u6237\u53ef\u4ee5\u5229\u7528\u4ee5\u4e0b\u547d\u4ee4\u4f7f\u7528\u6a21\u578b\u8fdb\u884c\u6458\u8981\u751f\u6210\uff0c\u6a21\u578b\u7684\u8def\u5f84\u7531\u2018checkpoint_dir\u2019\u6307\u5b9a\u3002\u7528\u6237\u53ef\u4ee5\u901a\u8fc7\u2018append_cols\u2019\u6307\u5b9a\u5728\u8f93\u51fa\u6587\u4ef6\u4e2d\u6dfb\u52a0\u8f93\u5165\u5217\uff0c\u5982\u679c\u4e0d\u6307\u5b9a\u5219\u586bnone\u3002python main.py \/<\/p>\n<pre><code>--mode=predict \/\n--app_name=sequence_generation \/\n--worker_gpu=1 \/\n--tables=.\/cn_dev.tsv  \/\n--outputs=.\/cn.preds.txt \/\n--input_schema=title:str:1,content:str:1,title_tokens:str:1,content_tokens:str:1,tag:str:1 \/\n--output_schema=predictions,beams \/\n--append_cols=content,title,tag \/\n--first_sequence=content_tokens \/\n--checkpoint_dir=.\/finetuned_zh_model \/\n--micro_batch_size=32 \/\n--sequence_length=512 \/\n--user_defined_parameters 'language=zh copy=false max_encoder_length=512 min_decoder_length=12 max_decoder_length=32 no_repeat_ngram_size=2 num_beams=5 num_return_sequences=5'<\/code><\/pre>\n<p>\u4ee5\u4e0b\u4e3a\u6a21\u578b\u5bf9\u8fd1\u671f\u70ed\u70b9\u4e8b\u4ef6\u9884\u6d4b\u7684\u51e0\u6761\u6837\u4f8b\uff0c\u6bcf\u6761\u6837\u4f8b\u5305\u542b5\u5217\u6570\u636e\uff08\u4ee5\u5236\u8868\u7b26\/t\u9694\u5f00\uff09\uff0c\u5206\u522b\u4e3a\u9884\u6d4b\u7684\u6458\u8981\u5217\uff08\u65b0\u95fb\u6807\u9898\uff09\u3001beam search\u76845\u6761\u5019\u9009\uff08\u7528||\u9694\u5f00\uff09\u3001\u8f93\u5165\u7684\u539f\u6587\u3001\u8f93\u5165\u7684\u65b0\u95fb\u6807\u7b7e\u3002\u5176\u4e2d\u540e\u4e09\u5217\u662f\u4ece\u5bf9\u5e94\u7684\u8f93\u5165\u6570\u636e\u4e2d\u76f4\u63a5\u62f7\u8d1d\u8fc7\u6765\u3002\u7531\u4e8e\u65b0\u95fb\u6587\u672c\u8fc7\u957f\uff0c\u4ee5\u4e0b\u4ec5\u5c55\u793a\u6bcf\u6761\u6837\u4f8b\u7684\u524d\u56db\u5217\u7ed3\u679c\u3002<strong>\u8d39\u5fb7\u52d2\u544a\u522b\u4fe1\uff1a\u672a\u6765\u6211\u8fd8\u4f1a\u6253\u66f4\u591a\u7684\u7f51\u7403<\/strong>  \u8d39\u5fb7\u52d2\u544a\u522b\u4fe1\uff1a\u672a\u6765\u6211\u8fd8\u4f1a\u6253\u66f4\u591a\u7684\u7f51\u7403||\u8d39\u5fb7\u52d2\u544a\u522b\u4fe1\uff1a\u672a\u6765\u6211\u8fd8\u4f1a\u6253\u66f4\u591a\u7f51\u7403||\u8d39\u5fb7\u52d2\u544a\u522b\u4fe1\uff1a\u672a\u6765\u6211\u8fd8\u4f1a\u6253\u66f4\u591a\u7f51\u7403\u4f46\u4e0d\u662f\u5728\u5927\u6ee1\u8d2f\u6216\u5de1\u56de\u8d5b||\u8d39\u5fb7\u52d2\u544a\u522b\u4fe1\uff1a\u672a\u6765\u6211\u8fd8\u4f1a\u6253\u66f4\u591a\u7684\u7f51\u7403||\u8be6\u8baf\uff1a\u8d39\u5fb7\u52d2\u5ba3\u5e03\u9000\u5f79\uff0c\u5e76\u544a\u522b\u4fe1  <strong>\u4e00\u4ee3\u4f20\u5947\u843d\u5e55\uff01\u7f51\u7403\u5929\u738b\u8d39\u5fb7\u52d2\u5ba3\u5e03\u9000\u5f79<\/strong>  \u592e\u89c6\u7f51\u6d88\u606f\uff1a\u5317\u4eac\u65f6\u95f49\u670815\u65e5\u665a\uff0c\u7f51\u7403\u5929\u738b\u7f57\u6770-\u8d39\u5fb7\u52d2\u5728\u4e2a\u4eba\u793e\u5a92\u4e0a\u5ba3\u5e03\u9000\u5f79\u300241\u5c81\u7684\u8d39\u5fb7\u52d2\u662f\u7537\u5b50\u7f51\u575b\u5386\u53f2\u6700\u4f1f\u5927\u7403\u5458\u4e4b\u4e00\uff0c\u66fe103\u6b21\u65a9\u83b7\u5355\u6253\u51a0\u519b\uff0c\u5927\u6ee1\u8d2f\u5355\u6253\u593a\u51a020\u6b21\uff08\u6fb3\u7f516\u51a0\u3001\u6cd5\u7f511\u51a0\u3001\u6e29\u7f518\u51a0\u3001\u7f8e\u7f515\u51a0\uff09\uff0c\u5171\u8ba1310\u5468\u4f4d\u4e8e\u7537\u5355\u4e16\u754c\u7b2c\u4e00\u3002\u9644\u8d39\u5fb7\u52d2\u544a\u522b\u4fe1\uff1a\u5728\u8fd9\u4e9b\u5e74\u7f51\u7403\u7ed9\u6211\u7684\u6240\u6709\u793c\u7269\u4e2d\uff0c\u6700\u68d2\u7684\u6beb\u65e0\u7591\u95ee\u662f\u6211\u4e00\u8def\u4e0a\u6240\u9047\u5230\u7684\u4eba\uff1a\u6211\u7684\u670b\u53cb\u3001\u6211\u7684\u7ade\u4e89\u5bf9\u624b\u3001\u4ee5\u53ca\u6700\u91cd\u8981\u7684\u7403\u8ff7\uff0c\u662f\u4ed6\u4eec\u7ed9\u4e88\u4e86\u8fd9\u9879\u8fd0\u52a8\u751f\u547d\u3002\u4eca\u5929\uff0c\u6211\u60f3\u548c\u5927\u5bb6\u5206\u4eab\u4e00\u4e9b\u6d88\u606f\u3002\u6b63\u5982\u4f60\u4eec\u4e2d\u7684\u8bb8\u591a\u4eba\u6240\u77e5\u9053\u7684\uff0c\u8fc7\u53bb\u4e09\u5e74\u4e2d\uff0c\u6211\u9047\u5230\u4e86\u53d7\u4f24\u548c\u624b\u672f\u7684\u6311\u6218\u3002&#8230;&#8230;<br \/><strong>\u53f0\u98ce\u201c\u6885\u82b1\u201d\u5c06\u5728\u5927\u8fde\u6cbf\u6d77\u767b\u9646\u5c06\u9010\u6b65\u53d8\u6027\u4e3a\u6e29\u5e26\u6c14\u65cb<\/strong>  \u53f0\u98ce\u201c\u6885\u82b1\u201d\u5c06\u5728\u5927\u8fde\u6cbf\u6d77\u767b\u9646\u5c06\u9010\u6b65\u53d8\u6027\u4e3a\u6e29\u5e26\u6c14\u65cb||\u53f0\u98ce\u201c\u6885\u82b1\u201d\u5c06\u5728\u5927\u8fde\u6cbf\u6d77\u767b\u9646\u540e\u9010\u6e10\u53d8\u6027\u4e3a\u6e29\u5e26\u6c14\u65cb||\u53f0\u98ce\u201c\u6885\u82b1\u201d\u5c06\u5728\u5927\u8fde\u6cbf\u6d77\u767b\u9646\u5c06\u9010\u6e10\u53d8\u6027\u4e3a\u6e29\u5e26\u6c14\u65cb||\u53f0\u98ce\u201c\u6885\u82b1\u201d\u5c06\u5728\u5927\u8fde\u6cbf\u6d77\u767b\u9646\u540e\u53d8\u6027\u4e3a\u6e29\u5e26\u6c14\u65cb||\u53f0\u98ce\u201c\u6885\u82b1\u201d\u5c06\u5728\u5927\u8fde\u6cbf\u6d77\u767b\u9646\u540e\u9010\u6e10\u53d8\u6027 <strong>\u53f0\u98ce\u201c\u6885\u82b1\u201d\u5c06\u4e8e16\u65e5\u508d\u665a\u524d\u540e\u5728\u8fbd\u5b81\u5927\u8fde\u6cbf\u6d77\u767b\u9646<\/strong>  \u8bb0\u80059\u670816\u65e5\u4ece\u8fbd\u5b81\u7701\u5927\u8fde\u5e02\u6c14\u8c61\u90e8\u95e8\u83b7\u6089\uff0c\u4eca\u5e74\u7b2c12\u53f7\u53f0\u98ce\u201c\u6885\u82b1\u201d\u5c06\u4e8e16\u65e5\u508d\u665a\u524d\u540e\u5728\u5927\u8fde\u5e02\u65c5\u987a\u53e3\u533a\u81f3\u5e84\u6cb3\u5e02\u4e00\u5e26\u6cbf\u6d77\u767b\u9646\uff0c\u4e4b\u540e\u9010\u6e10\u53d8\u6027\u4e3a\u6e29\u5e26\u6c14\u65cb\u3002\u3000\u3000\u53d7\u53f0\u98ce\u201c\u6885\u82b1\u201d\u5f71\u54cd\uff0c14\u65e58\u65f6\u81f316\u65e510\u65f6\uff0c\u5927\u8fde\u5168\u5e02\u5e73\u5747\u964d\u96e8\u91cf\u4e3a132\u6beb\u7c73\uff0c\u6700\u5927\u964d\u96e8\u91cf\u51fa\u73b0\u5728\u91d1\u666e\u65b0\u533a\u5927\u674e\u5bb6\u8857\u9053\u6b63\u660e\u5bfa\u6751\uff0c\u4e3a283.6\u6beb\u7c73\uff1b\u4e00\u5c0f\u65f6\u6700\u5927\u964d\u96e8\u91cf\u51fa\u73b0\u5728\u957f\u6d77\u53bf\u5e7f\u9e7f\u5c9b\u9547\uff0c\u4e3a49.4\u6beb\u7c73&#8230;&#8230;\u00a0\u82f1\u6587\u6587\u672c\u6458\u8981EasyNLP\u6a21\u578b\u5e93\u4e2d\u540c\u6837\u96c6\u6210\u4e86\u82f1\u6587\u6587\u672c\u6458\u8981\u6a21\u578b\uff0c\u5305\u62ecPEGASUS\u548cBRIO\u3002\u4ee5\u4e0b\u8868\u683c\u5c55\u793a\u4e86\u4e24\u4e2a\u6a21\u578b\u5728\u82f1\u6587\u6587\u672c\u6458\u8981\u6570\u636e\u4e0a\u7684\u6027\u80fd\u8868\u73b0\u3002\u7528\u6237\u540c\u6837\u53ef\u4ee5\u4f7f\u7528\u4e0a\u8ff0\u4ee3\u7801\u5bf9\u6a21\u578b\u8fdb\u884c\u8bad\u7ec3\u548c\u9884\u6d4b\u3002\u9700\u8981\u6ce8\u610f\u7684\u662f\uff0cEasyNLP\u9ed8\u8ba4\u7684\u662f\u5bf9\u4e2d\u6587\u7684\u5904\u7406\uff0c\u56e0\u6b64\uff0c\u5f53\u9700\u8981\u5904\u7406\u82f1\u6587\u6587\u672c\u65f6\uff0c\u9700\u8981\u5728\u2018user_defined_parameters\u2019\u4e2d\u6307\u5b9alanguage\u4e3aen\uff0c\u5982\u4e0d\u63d0\u4f9b\uff0c\u5219\u9ed8\u8ba4\u4e3a\u4e2d\u6587\uff08zh\uff09\u3002\u82f1\u6587\u6587\u672c\u6458\u8981\uff08Rouge1\/2\/L\uff09alibaba-pai\/pegasus-summary-generation-en37.79\/18.69\/35.44hfl\/brio-cnndm-uncased41.46\/23.34\/38.91\u00a0\u00a0\u8bad\u7ec3\u8fc7\u7a0b\u5982\u4e0b\uff1awget <a href=\"https:\/\/link.segmentfault.com\/?enc=%2BLO4LVfMI9lj2OwAm%2FbzNw%3D%3D.O%2FU3x%2B7UHHG7K5CpTqV%2FxxRez0flvtTSkzxR71pQK3HblRNUMtOQboqJW9cvphikBiNKUM9%2FlXdRaFMDbLoX7PUOm6BECtsQ1%2Br8Z8ywmmpDU1sg4VjaxbNuAHh6cBog\" rel=\"nofollow\">http:\/\/atp-modelzoo-sh.oss-cn&#8230;<\/a><br \/>wget <a href=\"https:\/\/link.segmentfault.com\/?enc=RLzNCEopVQ3fuBHYW%2FHBzw%3D%3D.%2FPAqhnkPJQhlm39bxZDgY6Kr%2FmD7vrTngqcSHrKT3WeTHmilLVrp4F5kkMR19XfsMx4%2F2FKkSiPLhFPM0drGfFjpbnYPIEmt%2FkjY0DmQlc5s7XDhCajUdZEl%2BnuauosW\" rel=\"nofollow\">http:\/\/atp-modelzoo-sh.oss-cn&#8230;<\/a><\/p>\n<p>python main.py \/<\/p>\n<pre><code>--mode train \/\n--app_name=sequence_generation \/\n--worker_gpu=1 \/\n--tables=.\/en_train.tsv,.\/en_dev.tsv  \/\n--input_schema=title:str:1,content:str:1 \/\n--first_sequence=content \/\n--second_sequence=title \/\n--label_name=title \/\n--checkpoint_dir=.\/finetuned_en_model \/\n--micro_batch_size=1 \/\n--sequence_length=512 \/\n--epoch_num 1 \/\n--save_checkpoint_steps=500 \/\n--export_tf_checkpoint_type none \/\n--user_defined_parameters 'language=en pretrain_model_name_or_path=alibaba-pai\/pegasus-summary-generation-en copy=false max_encoder_length=512 min_decoder_length=64 max_decoder_length=128 no_repeat_ngram_size=2 num_beams=5 num_return_sequences=5'<\/code><\/pre>\n<p>\u00a0\u9884\u6d4b\u8fc7\u7a0b\u5982\u4e0b\uff1apython main.py \/<\/p>\n<pre><code>--mode=predict \/\n--app_name=sequence_generation \/\n--worker_gpu=1 \/\n--tables=.\/en_dev.tsv  \/\n--outputs=.\/en.preds.txt \/\n--input_schema=title:str:1,content:str:1 \/\n--output_schema=predictions,beams \/\n--append_cols=title,content \/\n--first_sequence=content \/\n--checkpoint_dir=.\/finetuned_en_model \/\n--micro_batch_size 32 \/\n--sequence_length 512 \/\n--user_defined_parameters 'language=en copy=false max_encoder_length=512 min_decoder_length=64 max_decoder_length=128 no_repeat_ngram_size=2 num_beams=5 num_return_sequences=5'<\/code><\/pre>\n<p>\u00a0\u4ee5\u4e0b\u5c55\u793a\u4e86\u6a21\u578b\u5bf9\u4e00\u7bc7\u70ed\u70b9\u79d1\u6280\u65b0\u95fb\u7a3f\u7684\u6458\u8981\u9884\u6d4b\u7ed3\u679c\uff1aWith the image generator Stable Diffusion, you can conjure within seconds a potrait of Beyonc\u00e9 as if painted by Vincent van Gogh, a cyberpunk cityscape in the style of 18th century Japanese artist Hokusai and a complex alien world straight out of science fiction. Released to the public just two weeks ago, it\u2019s become one of several popular AI-powered text-to-image generators, including DALL-E 2, that have taken the internet by storm. Now, the company behind Stable Diffusion is in discussions to raise $100 million from investors, according to three people with knowledge of the matter. Investment firm Coatue expressed initial interest in a deal that would value the London-based startup Stability AI at $500 million, according to two of the people. Lightspeed Venture Partners then entered talks \u2014 which are still underway \u2014 to invest at a valuation up to $1 billion, two sources said. Stability AI, Coatue and Lightspeed declined requests for comment. The London-based startup previously raised at least $10 million in SAFE notes (a form of convertible security popular among early-stage startups) at a valuation of up to $100 million, according to one of the sources. An additional fourth source with direct knowledge confirmed Stability AI\u2019s previous round. Much of the company\u2019s funds came directly from founder and CEO Emad Mostaque, a former hedge fund manager. News of the prior financing was previously unreported. By nature of being open source, Stability AI\u2019s underlying technology is free to use. So far, the company does not have a clear business model in place, according to three of the sources. However, Mostaque said in an interview last month with Yannic Kilcher, a machine learning engineer and YouTube personality, that he has already penned partnerships with \u201cgovernments and leading institutions\u201d to sell the technology. \u201cWe\u2019ve negotiated massive deals so we\u2019d be profitable at the door versus most money-losing big corporations,\u201d he claims. The first version of Stable Diffusion itself cost just $600,000 to train, he wrote on Twitter \u2014 a fraction of the company\u2019s total funding. Mostaque, 39, hails from Bangladesh and grew up in England. He received a master\u2019s degree in mathematics and computer science from Oxford University in 2005 and spent 13 years working at U.K. hedge funds. In 2019, he launched Symmitree, a startup that aimed to reduce the cost of technology for people in poverty; it shuttered after one year, according to his LinkedIn profile. He then founded Stability AI in late 2020 with the mission of building open-source AI projects. According to its website, text-to-image generation is only one component of a broader apparatus of AI-powered offerings that the company is helping to build. Other open-source research groups it backs are developing tools for language, audio and biology. Stable Diffusion \u2014 created in collaboration with RunwayML, a video editing startup also backed by Coatue, and researchers at the Ludwig Maximilian University of Munich \u2014 has generated by far the most buzz among the company\u2019s projects. It comes as AI image generators entered the zeitgeist this year, with the release of OpenAI\u2019s DALL-E 2 in April and independent research lab Midjourney\u2019s eponymous product in July. Google also revealed a text-to-image system, Imagen, in May, though it is not available to the public. Mostaque and his peers have said that the existing technology only represents the tip of the iceberg of what AI art is capable of creating: Future use cases could include drastically improved photorealism, video and animation. These image generators are already facing controversy: Many of them have been trained by processing billions of images on the internet without the consent of the copyright holder, prompting debate over ethics and legality. Last week, a testy debate broke out online after a Colorado fine arts competition awarded a top prize to an AI-generated work of art. Moreover, unlike DALL-E and Midjourney, which have restrictions in place to prevent the generation of gory or pornographic images, Stable Diffusion\u2019s open source nature allows users to bypass such a block. On 4chan, numerous threads have appeared with AI-generated deepfakes of celebrity nudes, while Reddit has banned at least four communities that were dedicated to posting \u201cnot safe for work\u201d AI imagery made using Stable Diffusion. It\u2019s a double-edged sword for Stability AI, which has accumulated community goodwill precisely due to its open source approach that gives its users full access to its code. The company\u2019s website states that the company is \u201cbuilding open AI tools,\u201d a mission that mirrors the initial intent of OpenAI to democratize access to artificial intelligence. OpenAI was launched as a nonprofit research organization by prominent technologists including Sam Altman and Elon Musk, but upon accepting a $1 billion investment from Microsoft in 2019, it became a for-profit business. The move led it to focus on commercializing its technology rather than making it more widely available, drawing criticism from the AI community \u2014 and Musk himself.  Stability AI has been a for-profit corporation from its inception, which Mostaque has said is meant to allow the open source research to reach more people. In an interviewwith TechCrunch last month, he said that the company was fully independent. \u201cNobody has any voting rights except our 75 employees \u2014 no billionaires, big funds, governments or anyone else with control of the company or the communities we support,\u201d he said. At a $1 billion valuation, Mostaque would be ceding up to 10% of the company to the new financiers. Venture capital investors who take significant stakes in startups typically ask for board positions so they can influence the decisions the company is making using their money. Lightspeed, which manages $10 billion of assets, and Coatue, which is in charge of $73 billion, both have a track record of taking board seats, though it\u2019s unclear if that will be the case with Stability AI. Follow me on Twitter. Send me a secure tip. <br \/>\u4e0a\u8ff0\u6587\u672c\u6765\u81ea\u4e8e<a href=\"https:\/\/link.segmentfault.com\/?enc=7dfPsYoQDcQCXzOBQh2X3A%3D%3D.ivntzN3DPahX0W7PB%2FIoeH713QTzFtNGqyTKcLhBBU5ahBK10HaCOybhtxLOsHJNnsDZxndlfz6YTs4o%2F1b5J33ZzSOEIURqNSypToZqb9lryte5P%2BopX9I7gookMttrsqeqBvNU5rrBOMnW9m3wrUVb6%2FTh6EJlceymYtyeCjYBXTA9HzDWFsjCr4%2BFujoc1QsDV8isBnAiAm92PHxqhQ%3D%3D\" rel=\"nofollow\">https:\/\/www.forbes.com\/sites\/&#8230;<\/a>\u9488\u5bf9\u4e0a\u8ff0\u65b0\u95fb\u539f\u7a3f\uff0c\u4ee5\u4e0b\u4e3a\u4e24\u4e2a\u6700\u65b0\u6a21\u578b\u7684\u6458\u8981\u751f\u6210\u7ed3\u679c\uff1astable Diffusion is in discussions to raise $100 million from investors, three people say. The image generator is one of several popular AI-powered text-to-image generators.company behind the popular image generator Stable Diffusion is in talks to raise $100 million from investors, according to sources\u4ee5\u4e0a\u662f\u5982\u4f55\u5229\u7528EasyNLP\u8fdb\u884c\u6587\u672c\u6458\u8981\u6a21\u578b\u8bad\u7ec3\u548c\u9884\u6d4b\u7684\u5168\u90e8\u8fc7\u7a0b\uff0c\u66f4\u8be6\u7ec6\u7684\u4f7f\u7528\u6559\u7a0b\u53ef\u52a0\u5165\u4ee5\u4e0b\u8bfe\u7a0b\u8fdb\u884c\u5b66\u4e60\u3002\u6807\u9898\u515a\u901f\u6210\u73ed\uff1a\u57fa\u4e8e\u673a\u5668\u5b66\u4e60PAI EasyNLP\u7684\u4e2d\u6587\u65b0\u95fb\u6807\u9898\u751f\u6210\u672a\u6765\u5c55\u671b\u5728\u672a\u6765\uff0c\u6211\u4eec\u8ba1\u5212\u5728EasyNLP\u6846\u67b6\u4e2d\u96c6\u6210\u9762\u5411\u77e5\u8bc6\u7684\u4e2d\u2f42\u9884\u8bad\u7ec3\u6a21\u578b\uff0c\u8986\u76d6\u5404\u4e2a\u5e38\u2ec5\u7684NLU\u548cNLG\u4e2d\u2f42\u9886\u57df\uff0c\u656c\u8bf7\u671f\u5f85\u3002\u6211\u4eec\u4e5f\u5c06\u5728EasyNLP\u6846\u67b6\u4e2d\u96c6\u6210\u66f4\u591aSOTA\u6a21\u578b\uff08\u7279\u522b\u662f\u4e2d\u2f42\u6a21\u578b\uff09\uff0c\u6765\u2f40\u6301\u5404\u79cdNLP\u548c\u591a\u6a21\u6001\u4efb\u52a1\u3002\u6b64\u5916\uff0c \u963f\u2fa5\u4e91\u673a\u5668\u5b66\u4e60PAI\u56e2\u961f\u4e5f\u5728\u6301\u7eed\u63a8\u8fdb\u4e2d\u6587\u6587\u672c\u751f\u6210\u548c\u4e2d\u2f42\u591a\u6a21\u6001\u6a21\u578b\u7684\u2f83\u7814\u2f2f\u4f5c\uff0c\u6b22\u8fce\u2f64\u6237\u6301\u7eed\u5173\u6ce8\u6211\u4eec\uff0c\u4e5f\u6b22\u8fce\u52a0\u2f0a \u6211\u4eec\u7684\u5f00\u6e90\u793e\u533a\uff0c\u5171\u5efa\u4e2d\u2f42NLP\u548c\u591a\u6a21\u6001\u7b97\u6cd5\u5e93\uff01\u00a0Github\u5730\u5740\uff1a<a href=\"https:\/\/link.segmentfault.com\/?enc=ctdZkiC2mcMfm8KpcMMFQQ%3D%3D.BDqHtKs8F7wXv97EgiJD%2B6c3wjallgmvcaItxosdVuu9Eu8tKqTE7qqTS3L7LQWV\" rel=\"nofollow\">https:\/\/github.com\/alibaba\/Ea&#8230;<\/a>\u00a0\u53c2\u8003\u6587\u732eChengyu Wang, Minghui Qiu, Taolin Zhang, Tingting Liu, Lei Li, Jianing Wang, Ming Wang, Jun Huang, Wei Lin. EasyNLP: A Comprehensive and Easy-to-use Toolkit for Natural Language Processing. arXivZhang, Jingqing, et al. &#8220;Pegasus: Pre-training with extracted gap-sentences for abstractive summarization.&#8221;\u00a0International Conference on Machine Learning. PMLR, 2020.Xue, Linting, et al. &#8220;mT5: A massively multilingual pre-trained text-to-text transformer.&#8221;\u00a0arXiv preprint arXiv:2010.11934(2020).Lewis, Mike, et al. &#8220;Bart: Denoising sequence-to-sequence pre-training for natural language generation, translation, and comprehension.&#8221;\u00a0arXiv preprint arXiv:1910.13461\u00a0(2019).Song, Kaitao, et al. &#8220;Mass: Masked sequence to sequence pre-training for language generation.&#8221;\u00a0arXiv preprint arXiv:1905.02450\u00a0(2019).Dong, Li, et al. &#8220;Unified language model pre-training for natural language understanding and generation.&#8221;\u00a0Advances in Neural Information Processing Systems\u00a032 (2019).Yixin Liu, Pengfei Liu, Dragomir Radev, and Graham Neubig. 2022.\u00a0BRIO: Bringing Order to Abstractive Summarization. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 2890\u20132903, Dublin, Ireland. Association for Computational Linguistics.\u00a0\u963f\u91cc\u7075\u6770\u56de\u987e\u963f\u91cc\u7075\u6770\uff1a\u963f\u91cc\u4e91\u673a\u5668\u5b66\u4e60PAI\u5f00\u6e90\u4e2d\u6587NLP\u7b97\u6cd5\u6846\u67b6EasyNLP\uff0c\u52a9\u529bNLP\u5927\u6a21\u578b\u843d\u5730\u963f\u91cc\u7075\u6770\uff1a\u9884\u8bad\u7ec3\u77e5\u8bc6\u5ea6\u91cf\u6bd4\u8d5b\u593a\u51a0\uff01\u963f\u91cc\u4e91PAI\u53d1\u5e03\u77e5\u8bc6\u9884\u8bad\u7ec3\u5de5\u5177\u963f\u91cc\u7075\u6770\uff1aEasyNLP\u5e26\u4f60\u73a9\u8f6cCLIP\u56fe\u6587\u68c0\u7d22\u963f\u91cc\u7075\u6770\uff1aEasyNLP\u4e2d\u6587\u6587\u56fe\u751f\u6210\u6a21\u578b\u5e26\u4f60\u79d2\u53d8\u827a\u672f\u5bb6\u963f\u91cc\u7075\u6770\uff1aEasyNLP\u96c6\u6210K-BERT\u7b97\u6cd5\uff0c\u501f\u52a9\u77e5\u8bc6\u56fe\u8c31\u5b9e\u73b0\u66f4\u4f18Finetune\u539f\u6587\u94fe\u63a5\uff1a<a href=\"https:\/\/link.segmentfault.com\/?enc=a%2FZ9%2FdHt4svuHHm9xkfDjQ%3D%3D.mCd02cw%2BiRSXaEWsnTs6nm2bFHl4YwKgsep%2BbWiX8TPqrIm5UWptbyJFo%2BmjKZhw\" rel=\"nofollow\">https:\/\/click.aliyun.com\/m\/10&#8230;<\/a>\u672c\u6587\u4e3a\u963f\u91cc\u4e91\u539f\u521b\u5185\u5bb9\uff0c\u672a\u7ecf\u5141\u8bb8\u4e0d\u5f97\u8f6c\u8f7d\u3002<\/p>\n<p><a href=\"https:\/\/segmentfault.com\/a\/1190000042566821\">https:\/\/segmentfault.com\/a\/1190000042566821<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u7b80\u4ecb\uff1a\u00a0\u672c\u2f42\u5c06\u63d0\u4f9b\u5173\u4e8ePEGASUS\u7684\u6280\u672f\u89e3\u8bfb\uff0c\u4ee5\u53ca\u5982\u4f55\u5728EasyNLP\u6846\u67b6\u4e2d\u4f7f\u2f64\u4e0ePEGASUS\u76f8\u5173\u7684\u6587\u672c\u6458 [&hellip;]<\/p>\n","protected":false},"author":0,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[],"tags":[],"class_list":["post-10006641","post","type-post","status-publish","format-standard","hentry"],"jetpack_sharing_enabled":true,"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/posts\/10006641","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/itmax.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=10006641"}],"version-history":[{"count":0,"href":"https:\/\/itmax.cn\/index.php?rest_route=\/wp\/v2\/posts\/10006641\/revisions"}],"wp:attachment":[{"href":"https:\/\/itmax.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=10006641"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/itmax.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=10006641"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/itmax.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=10006641"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}