请问能开放基于clue数据的base模型，large太大训不动，tiny又不如bert base。

[InsaneLife]

如题。 因为看模型对比里面有BERT-base基于clue_vocab训练的，做了速度对比。

ps: RoBERTa-large-clue词表和配置不一样。vocab大小是8021，config文件大小是21128。

[DukeEnglish]

预训练模型的事情我们会在后续规划。词表的问题我们收到了，后续会更新。谢谢发自我的iPhone------------------ 原始邮件 ------------------发件人: AaronChou [email protected]发送时间: 2020年3月26日 21:50收件人: CLUEbenchmark/CLUEPretrainedModels [email protected]抄送: Subscribed [email protected]主题: 回复：[CLUEbenchmark/CLUEPretrainedModels] 请问能开放基于clue数据的base模型，large太大训不动，tiny又不如bert base。 (#11) 如题。 ps: RoBERTa-large-clue词表和配置不一样。vocab大小是8021，config文件大小是21128。

—You are receiving this because you are subscribed to this thread.Reply to this email directly, view it on GitHub, or unsubscribe. [ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/CLUEbenchmark/CLUEPretrainedModels/issues/11", "url": "https://github.com/CLUEbenchmark/CLUEPretrainedModels/issues/11", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

[InsaneLife]

因为看模型对比里面有BERT-base基于clue_vocab训练的，做了速度对比，能否开放更多模型，最好是base，太大训练不动啊