something about loss

[fanmanqian]

Thank you for sharing your code.By running your code,during the training period,the final loss in my own datasets：3358.2684.Why does the loss reach thousands? Is this normal?

[JusperLee]

This is not normal, because the loss function in my code is generally negative.

fanmanqian notifications@github.com 于2020年8月12日周三 上午9:33写道：

Thank you for sharing your code.By running your code,during the training period,the final loss in my own datasets：3358.2684.Why does the loss reach thousands? Is this normal?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/JusperLee/Deep-Clustering-for-Speech-Separation/issues/3, or unsubscribe https://github.com/notifications/unsubscribe-auth/AIB5NQXVD6EFOVNUIX275RLSAHWO7ANCNFSM4P33VLIA .

-- Kai Li Student, Tsinghua University

18594841812 | tsinghua.kaili@gmail.com

likai.show https://www.likai.show/ Tsinghua, Beijing, China

[fanmanqian]

ok,i thought this loss value looks like this picture like yours.so I put up this issue.

[fanmanqian]

我原本以为我的值和作者您的这个图差不多 然后就稍微问了一下这个问题，估计我的哪里出问题了，我再看看，谢谢您的回复~

[JusperLee]

我好像说错了，只要是你的loss呈下降趋势即可

[JusperLee]

因为他是针对于tf-bin进行计算的loss，因此它的loss都是比较大的

[fanmanqian]

好哒，谢谢您的热心回复~

[fanmanqian]

好哒，谢谢您~

------------------ 原始邮件 ------------------ 发件人: "Kai Li (李凯)"<notifications@github.com>; 发送时间: 2020年8月14日(星期五) 下午3:13 收件人: "JusperLee/Deep-Clustering-for-Speech-Separation"<Deep-Clustering-for-Speech-Separation@noreply.github.com>; 抄送: "孟凡满"<2563965485@qq.com>; "Author"<author@noreply.github.com>; 主题: Re: [JusperLee/Deep-Clustering-for-Speech-Separation] something about loss (#3)

因为他是针对于tf-bin进行计算的loss，因此它的loss都是比较大的

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.