2017년 10월 18일 수요일

dcGAN CNN architecture

kraken@devBox1:~/examples/dcgan$
kraken@devBox1:~/examples/dcgan$ python3 main.py --dataset folder --dataroot './data' --worker 16 --cuda --ngpu 2
Namespace(batchSize=16, beta1=0.5, cuda=True, dataroot='./data', dataset='folder', imageSize=64, lr=0.0002, manualSeed=None, ndf=64, netD='', netG='', ngf=64, ngpu=2, niter=3000, nz=100, outf='.', workers=16)
Random Seed:  1783
_netG (
  (main): Sequential (
    (0): ConvTranspose2d(100, 512, kernel_size=(4, 4), stride=(1, 1), bias=False)
    (1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True)
    (2): ReLU (inplace)
    (3): ConvTranspose2d(512, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (4): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True)
    (5): ReLU (inplace)
    (6): ConvTranspose2d(256, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (7): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True)
    (8): ReLU (inplace)
    (9): ConvTranspose2d(128, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (10): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True)
    (11): ReLU (inplace)
    (12): ConvTranspose2d(64, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (13): Tanh ()
  )
)
_netD (
  (main): Sequential (
    (0): Conv2d(3, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (1): LeakyReLU (0.2, inplace)
    (2): Conv2d(64, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (3): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True)
    (4): LeakyReLU (0.2, inplace)
    (5): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (6): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True)
    (7): LeakyReLU (0.2, inplace)
    (8): Conv2d(256, 512, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1), bias=False)
    (9): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True)
    (10): LeakyReLU (0.2, inplace)
    (11): Conv2d(512, 1, kernel_size=(4, 4), stride=(1, 1), bias=False)
    (12): Sigmoid ()
  )
)
[0/3000][0/57] Loss_D: 1.9780 Loss_G: 7.6137 D(x): 0.6910 D(G(z)): 0.7167 / 0.0007
[0/3000][1/57] Loss_D: 0.3106 Loss_G: 7.1012 D(x): 0.9215 D(G(z)): 0.1827 / 0.0011

댓글 없음:

댓글 쓰기

gpustat command

sudo apt install gpustat watch --color -n0.1 gpustat --color