더북(TheBook)

다음은 생성자 네트워크에 대한 출력 결과입니다.

Model: "sequential"
__________________________________________________________________
Layer (type)                 Output Shape                Param #
==================================================================
dense (Dense)                (None, 12544)               1266944
__________________________________________________________________
leaky_re_lu (LeakyReLU)      (None, 12544)               0
__________________________________________________________________
reshape (Reshape)            (None, 7, 7, 256)           0
__________________________________________________________________
conv2d_transpose (Conv2DTran (None, 7, 7, 128)           819328
__________________________________________________________________
batch_normalization (BatchNo (None, 7, 7, 128)           512
__________________________________________________________________
leaky_re_lu_1 (LeakyReLU)    (None, 7, 7, 128)           0
__________________________________________________________________
dropout (Dropout)            (None, 7, 7, 128)           0
__________________________________________________________________
conv2d_transpose_1 (Conv2DTr (None, 14, 14, 64)          204864
__________________________________________________________________
batch_normalization_1 (Batch (None, 14, 14, 64)          256
__________________________________________________________________
leaky_re_lu_2 (LeakyReLU)    (None, 14, 14, 64)          0
__________________________________________________________________
conv2d_transpose_2 (Conv2DTr (None, 28, 28, 1)           1601
==================================================================
Total params: 2,293,505
Trainable params: 2,293,121
Non-trainable params: 384
__________________________________________________________________

생성자 네트워크가 완료되었고, 이제 판별자 네트워크를 생성해 보겠습니다.

신간 소식 구독하기
뉴스레터에 가입하시고 이메일로 신간 소식을 받아 보세요.