ShaneTian

**You can use this:** ```py tokenizer = tf.keras.preprocessing.text.Tokenizer(oov_token="") tokenizer.fit_on_texts(x_text) x = tokenizer.texts_to_sequences(x_text) x = tf.keraspreprocessing.sequence.pad_sequences(x, maxlen=max_document_length, padding='post', truncating='post') ```

『在返回结果里添加意图和词槽』可以描述地更详细一些吗？或者举个例子

可以手动调用技能接口查看返回的提示信息在哪个字段，然后在 USKit 的 backend 中对应的 service 的 response 中配置返回

Hello~ Thanks for your attention on this work! We do not include the code about them. But in details, - the implementation of Speech Disfluency Simulator is similar as [LAUG](https://github.com/thu-coai/LAUG/tree/main/LAUG/aug/Speech_Disfluency)...

Excellent 💥

目前我发现Python可以高亮，shell不行……

@codevvvv9 在这里下载一个github图标http://www.iconfont.cn/ 放到/yelee/source/img/下命名为GitHub.png 在/yelee/source/css/_partial/customise/social-icon.styl的img-logo中加入GitHub black 75 删掉46-50行的.GitHub background url(//cdn.bootcss.com/logos/0.2.0/github-octocat.svg) no-repeat white background-size 90% background-position 50% 100% 即可 见这里 [GitHub图标设置后不能显示](https://suixinblog.cn/2018/09/hexo.html#GitHub%E5%9B%BE%E6%A0%87%E8%AE%BE%E7%BD%AE%E5%90%8E%E4%B8%8D%E8%83%BD%E6%98%BE%E7%A4%BA)

好像是5个就自动换行，可以在主题的css中找找

> > And please upgrade to 4.36.2 > > This problem occurs in training with multiple machines and multiple cards. Perhaps 4.36.2 did not solve this problem either, as 4.36.1...

@DesperateL 对，Valine比较好用。支持匿名、自定义头像、显示浏览器系统