search

carpedm20 / DCGAN-tensorflow

A tensorflow implementation of "Deep Convolutional Generative Adversarial Networks"
http://carpedm20.github.io/faces/
MIT License
7.14k stars 2.63k forks source link

Failed to generate test images more than 100 #349

Open RuikangSun opened 5 years ago

RuikangSun commented 5 years ago

python main.py --dataset dadada --input_height 64 --input_width 64 --output_height 64 --output_width 64 --visualize True --batch_size=1 --option=1 --generate_test_images 1000

RuikangSun commented 5 years ago

I followed #216 but it didn't work.

RuikangSun commented 5 years ago

No matter how I changed "--generate_test_images",it only generates 100 images.

ALK416 commented 5 years ago

No matter how I changed "--generate_test_images",it only generates 100 images.

Hi, I think I've seen this in other issues, if you want to generate more pic in train process. The command --sample_freq may help. if you don't set sample_freq it will be 200, and it means generate a pic every 200 iterations.

See the code in main.py : flags.DEFINE_integer("sample_freq", 200, "sample every this many iterations")

RuikangSun commented 5 years ago

No matter how I changed "--generate_test_images",it only generates 100 images.

Hi, I think I've seen this in other issues, if you want to generate more pic in train process. The command --sample_freq may help. if you don't set sample_freq it will be 200, and it means generate a pic every 200 iterations.

See the code in main.py : flags.DEFINE_integer("sample_freq", 200, "sample every this many iterations")

Thanks for replying! However, I want to generate more images in test process. ; )