General Code review on 3D model about focal-tversky-unet

[quqixun]

conv10 = Conv3D(3, (1, 1, 1), activation='relu', kernel_initializer=kinit,padding = 'same', name='final',
data_format = 'channels_first')(conv9)
activation_name = 'sigmoid'
activation_block = Activation(activation_name)(conv10)

Your output has 3 channels, you may apply 'softmax' to compute probability for each class.

[DecentMakeover]

Hi @quqixun ,Thanks for the reply
Changing activation from sigmoid to softmax does not work.

[nabsabraham]

Hi @DecentMakeover, did you find a solution? I actually haven't tried 3D but will run your code sometime this week to see if I can debug

[DecentMakeover]

Hey , that’s okay , I have moved on from this now. On 02-Aug-2019, at 8:47 PM, Nabila Abraham <[email protected]<mailto:[email protected]>> wrote: Hi @DecentMakeover<https://github.com/DecentMakeover>, did you find a solution? I actually haven't tried 3D but will run your code sometime this week to see if I can debug — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#8?email_source=notifications&email_token=AGD5QF6FNVOPVJ63T2ECWOTQCRFXXA5CNFSM4H2WVJTKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3OBJEQ#issuecomment-517739666>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AGD5QF3BB2AXHKLEJCUZ3XDQCRFXXANCNFSM4H2WVJTA>.