search

tornadomeet / mxnet-face

Using mxnet for face-related algorithm.
Apache License 2.0
544 stars 206 forks source link

Change Moonoutput to SoftmaxOutput #26

Closed charleslipku closed 7 years ago

charleslipku commented 7 years ago

@tornadomeet I try to train my own model using SoftmaxOutput instead of MoonOutput. Can I still use the same network architecture and similar hyper parameters? Could you give me some tips on this? Since I tried several times, but he prediction is always NaN.

tornadomeet commented 7 years ago

SoftmaxOutput do not support your problem.

charleslipku commented 7 years ago

Then what other options are suitable for my problem instead of MoonOutput? How about LogisticrRegression?

On September 25, 2016 at 11:52:32 PM, Wei Wu (notifications@github.com) wrote:

SoftmaxOutput do not support your problem.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/tornadomeet/mxnet-face/issues/26#issuecomment-249479004, or mute the thread https://github.com/notifications/unsubscribe-auth/AOfH9tJFGqE9g8c2hOfJcwiugRGkyIjJks5qt0-QgaJpZM4KGISr .

charleslipku commented 7 years ago

Could you give me some suggestions about my problem? Thanks.

On September 26, 2016 at 12:08:24 AM, ANG LI (charleslipku@gmail.com) wrote:

Then what other options are suitable for my problem instead of MoonOutput? How about LogisticrRegression?

On September 25, 2016 at 11:52:32 PM, Wei Wu (notifications@github.com) wrote:

SoftmaxOutput do not support your problem.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/tornadomeet/mxnet-face/issues/26#issuecomment-249479004, or mute the thread https://github.com/notifications/unsubscribe-auth/AOfH9tJFGqE9g8c2hOfJcwiugRGkyIjJks5qt0-QgaJpZM4KGISr .

tornadomeet commented 7 years ago

sorry, i don't know.