ladder net g_guass unclear

[thebeancounter]

Can't find any reference to that formula in call and in g_guass in the paper and online, can you explain the origin of the formula

mu = a1 * tf.sigmoid(a2 * u + a3) + a4 * u + a5
v = a6 * tf.sigmoid(a7 * u + a8) + a9 * u + a10

z_est = (z_c - mu) * v + mu
return z_est

[divamgupta]

Hi, the formula of the denoising function is given in the appendix of the paper : "B Formulation of the Denoising Function"

[divamgupta]

We compile the model with categorical cross entropy, hence that loss is added by keras while compiling.

…

On Fri, 13 Sep 2019, 22:07 jeffmu, ***@***.***> wrote: hi, about the cost, you just use the decoder cost(d_cost), however, the c_cost you dont use it? Right? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2?email_source=notifications&email_token=AAONR5ORN56E74NGCM4GZ63QJRBMPA5CNFSM4HYWCAV2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD6WRUHQ#issuecomment-531438110>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAONR5IXHNWUEB5PS2LJBGTQJRBMPANCNFSM4HYWCAVQ> .

[wuhaizhifeng]

Thanks， I got it.

…

------------------ Phd student: Mu Guoqing Institute of Advanced Control Technology School of Control Science & Engineering Dalian University of Technology (DLUT), P. R. China Email： guoqingmu@foxmail.com

------------------ Original ------------------ From: "Divam Gupta"<notifications@github.com>; Date: Tue, Sep 17, 2019 01:52 PM To: "divamgupta/ladder_network_keras"<ladder_network_keras@noreply.github.com>; Cc: "穆国庆"<guoqingmu@foxmail.com>;"Comment"<comment@noreply.github.com>; Subject: Re: [divamgupta/ladder_network_keras] ladder net g_guass unclear (#2) We compile the model with categorical cross entropy, hence that loss is added by keras while compiling. On Fri, 13 Sep 2019, 22:07 jeffmu, <notifications@github.com> wrote: hi, about the cost, you just use the decoder cost(d_cost), however, the c_cost you dont use it? Right? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2?email_source=notifications&email_token=AAONR5ORN56E74NGCM4GZ63QJRBMPA5CNFSM4HYWCAV2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD6WRUHQ#issuecomment-531438110>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAONR5IXHNWUEB5PS2LJBGTQJRBMPANCNFSM4HYWCAVQ> . — You are receiving this because you commented. Reply to this email directly, view it on GitHub, or mute the thread.