search

VITA-Group / EnlightenGAN

[IEEE TIP] "EnlightenGAN: Deep Light Enhancement without Paired Supervision" by Yifan Jiang, Xinyu Gong, Ding Liu, Yu Cheng, Chen Fang, Xiaohui Shen, Jianchao Yang, Pan Zhou, Zhangyang Wang
Other
893 stars 198 forks source link

The artifact problem #21

Closed zhengziqiang closed 4 years ago

zhengziqiang commented 4 years ago

I followed your instructions and the results have some artifacts, especially the clean images. image

yifanjiang19 commented 4 years ago

Is this result produced by our pretrained model? We do not observe these artifacts before.

zhengziqiang commented 4 years ago

I just train the model using the provided datasets using the default parameters except set batchsize 1. Could this be the cause of the problem?

yifanjiang19 commented 4 years ago

Batch size should be no less than 8

zhengziqiang commented 4 years ago

OK, I will close this issue. Thank you.