Zehao Huang
Zehao Huang
Hi, This code only provides the implementation of SRResNet. SRGAN is implemented but the result is not very good.
It's implemented, but I don't provide in this repo. 在 2017年4月30日，13:33，OmrySendik 写道： What do you mean it is implemented? I do not see the adversarial loss. — You are receiving...
Sorry for my misleading information. I didn't provide the part of gan since the performance wasn't very good. I will share the codes if I can implement a good result....
Hi, I don't have time to update this codes. Also, I didn't benchmark the trained model. You can benchmark it by yourself. The results of SRResNet can be easily reproduced,...
Have you update your codes? https://github.com/huangzehao/torch-srgan/blob/master/test.lua#L34 It is caused by batchnorm, so I set model:evaluate() to model:training().
Hi, sorry, I think this codes do not support multiple GPU. Some codes about data loader and training should be modified (I am not very sure since I am not...
Hi, thanks for your carefully review. (1) This implementation is based on the the [v2 version](https://arxiv.org/pdf/1609.04802v2.pdf) of paper srgan, and there is not skip-connection. I also have not time to...
Hi, there maybe a typo in the readme of PixelShuffle. You can not convert a tensor of shape [C * r, H, W] to a tensor of shape [C, H...
Can you train COCO sucessfully?
Can you provide your training config?