a easy problem

[KarsonYu]

The initial paper directly uses $\epsilon$ minus the noise estimated by model in the training stage,why the way you use divides by square root 1-$a_0$ additionally?

[neillu23]

Hi @KarsonYu, I am not sure about your question. Could you give me more information about the equations you mentioned? (like which equation in which paper)

[KarsonYu]

It is the denominator in the algorithm 1 - training,the way you use to compute the gradient descent.

[KarsonYu]

Hi @KarsonYu, I am not sure about your question. Could you give me more information about the equations you mentioned? (like which equation in which paper)

I get it!But now I have another question.How to deduce Eq.15?

[KarsonYu]

Hi @KarsonYu, I am not sure about your question. Could you give me more information about the equations you mentioned? (like which equation in which paper)

Actually I can't understand almost all Eqs in Section3.3,do you have the detailed derivation process or give me some mathematical suggestions to solve it?

[KarsonYu]

It is the denominator in the algorithm 1 - training,the way you use to compute the gradient descent.

BTW,I find the paramter delta[t] in Eq.(9) is different between your paper and code.It equals 1-(1+m_{t}^2)*\bar{\alpha}_{t} in your code.I wonder why?

[XJTUmuxin]

公式9这个应该只是把括号加在后面两项再提了一个 $\bar{\alpha}_{t}$，但是我不知道他这公式是怎么来的

[KarsonYu]

公式9这个应该只是把括号加在后面两项再提了一个 $\bar{\alpha}_{t}$，但是我不知道他这公式是怎么来的

你能复现出论文的结果吗

[XJTUmuxin]

我拿他train好的模型推理了几个噪声语音，都是他github上给出的例子，结果和他的结果确实差不多，但是我没在整个数据集上去评估   张沐薪 ***@***.***  

…

------------------ 原始邮件 ------------------ 发件人: "Yu ***@***.***>; 发送时间: 2023年5月31日(星期三) 下午5:35 收件人: ***@***.***>; 抄送: ***@***.***>; ***@***.***>; 主题: Re: [neillu23/CDiffuSE] a easy problem (Issue #8) 公式9这个应该只是把括号加在后面两项再提了一个 $\bar{\alpha}_{t}$，但是我不知道他这公式是怎么来的 你能复现出论文的结果吗 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: ***@***.***>

[KarsonYu]

我拿他train好的模型推理了几个噪声语音，都是他github上给出的例子，结果和他的结果确实差不多，但是我没在整个数据集上去评估   张沐薪 @.***  
…
------------------ 原始邮件 ------------------ 发件人: "Yu @.>; 发送时间: 2023年5月31日(星期三) 下午5:35 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [neillu23/CDiffuSE] a easy problem (Issue #8) 公式9这个应该只是把括号加在后面两项再提了一个 $\bar{\alpha}_{t}$，但是我不知道他这公式是怎么来的 你能复现出论文的结果吗 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

PESQ分数也差不多吗？我计算分数差的有点远

[XJTUmuxin]

这个我没去测，大致听了一下，我不是做语音的

[MichaelChen147]

你们好，我是研二学生，在研究基于扩散模型的语音增强，方便交流一下吗？

[KarsonYu]

你们好，我是研二学生，在研究基于扩散模型的语音增强，方便交流一下吗？

我就水个论文，论文发了已经不做了

[MichaelChen147]

那可以看下你的论文吗？我也想学习下 karson ***@***.***>于2024年2月12日 周一21:14写道：

…

你们好，我是研二学生，在研究基于扩散模型的语音增强，方便交流一下吗？ 我就水个论文，论文发了已经不做了 — Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA> . You are receiving this because you commented.Message ID: ***@***.***>

[KarsonYu]

那可以看下你的论文吗？我也想学习下 karson @.>于2024年2月12日 周一21:14写道：
…
你们好，我是研二学生，在研究基于扩散模型的语音增强，方便交流一下吗？ 我就水个论文，论文发了已经不做了 — Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA . You are receiving this because you commented.Message ID: @.>

还没发出来，投了出去

[MichaelChen147]

有人在做这个课题，我也会好受些，期待学习你的论文 karson ***@***.***>于2024年2月12日 周一21:18写道：

…

那可以看下你的论文吗？我也想学习下 karson *@*. *>于2024年2月12日 周一21:14写道： … <#m_-2875696340101971545_> 你们好，我是研二学生，在研究基于扩散模型的语音增强，方便交流一下吗？ 我就水个论文，论文发了已经不做了 — Reply to this email directly, view it on GitHub <#8 (comment) <#8 (comment)>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA <https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA> . You are receiving this because you commented.Message ID: @.*> 还没发出来，投了出去 — Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AH7FIPD73X2DOCSSA5W6TM3YTICBLAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3TENJXGU> . You are receiving this because you commented.Message ID: ***@***.***>

[KarsonYu]

有人在做这个课题，我也会好受些，期待学习你的论文 karson @.>于2024年2月12日 周一21:18写道：
…
那可以看下你的论文吗？我也想学习下 karson @. >于2024年2月12日 周一21:14写道： … <#m_-2875696340101971545_> 你们好，我是研二学生，在研究基于扩散模型的语音增强，方便交流一下吗？ 我就水个论文，论文发了已经不做了 — Reply to this email directly, view it on GitHub <#8 (comment) <#8 (comment)>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA . You are receiving this because you commented.Message ID: @.> 还没发出来，投了出去 — Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7FIPD73X2DOCSSA5W6TM3YTICBLAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3TENJXGU . You are receiving this because you commented.Message ID: @.>

我纯为了毕业要求，之前发SCI被拒了，改了个普通的期刊。应该后续还有人在做的，多关注关注ICASSP和INTERSPEECH

[MichaelChen147]

希望你投稿顺利呢 karson ***@***.***>于2024年2月12日 周一21:25写道：

…

有人在做这个课题，我也会好受些，期待学习你的论文 karson *@*. *>于2024年2月12日 周一21:18写道： … <#m_6607294757252791648_> 那可以看下你的论文吗？我也想学习下 karson @. >于2024年2月12日 周一21:14写道： … <#m_-2875696340101971545_> 你们好，我是研二学生，在研究基于扩散模型的语音增强，方便交流一下吗？ 我就水个论文，论文发了已经不做了 — Reply to this email directly, view it on GitHub <#8 <#8> (comment) <#8 (comment) <#8 (comment)>>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA <https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA> https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA <https://github.com/notifications/unsubscribe-auth/AH7FIPDTSDALOM5WOJPE3PLYTIBTFAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3DMMZXGA> . You are receiving this because you commented.Message ID: @.> 还没发出来，投了出去 — Reply to this email directly, view it on GitHub <#8 (comment) <#8 (comment)>>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7FIPD73X2DOCSSA5W6TM3YTICBLAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3TENJXGU <https://github.com/notifications/unsubscribe-auth/AH7FIPD73X2DOCSSA5W6TM3YTICBLAVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU3TENJXGU> . You are receiving this because you commented.Message ID: @.*> 我纯为了毕业要求，之前发SCI被拒了，改了个普通的期刊。应该后续还有人在做的，多关注关注ICASSP和INTERSPEECH — Reply to this email directly, view it on GitHub <#8 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AH7FIPCJ4DGIF42NHEDH6JDYTIC37AVCNFSM6AAAAAAS2YMRF2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZYGU4DGMZZGI> . You are receiving this because you commented.Message ID: ***@***.***>