ExpGy

About Rain200H dataset training dim problem?

5

Hi, I am using the Rain200H dataset with a resolution of 481*321 cropped to 128*128 input size, but I am having the following problem. 

can not install json and pprint

12


warnings.warn('You have chosen a specific GPU. This will completely ' Use GPU: 0 for training Loaded pretrained weights for efficientnet-b3 => using pre-trained model 'efficientnet-b3' Using image size 300 Traceback...

Progressive Learning with adaptive Regularization

1

What part of the code can reflect the Progressive Learning with adaptive Regularization?

SPA数据有6万多张真实的标签和输入的样本对，但是训练出来的结果比输入图像还差。请问这是为啥？

我的环境： pytorch==0.4.1 python==3.6.13 错误代码： def forward(self,input): return irnn()(input,self.up_weight.weight, self.right_weight.weight,self.down_weight.weight, self.left_weight.weight, self.up_weight.bias,self.right_weight.bias, self.down_weight.bias,self.left_weight.bias) 是不是这个版本不支持这么调用 我改成 o= irnn.forward(self,input,self.up_weight.weight, self.right_weight.weight,self.down_weight.weight, self.left_weight.weight, self.up_weight.bias,self.right_weight.bias, self.down_weight.bias,self.left_weight.bias AttributeError: 'Spacial_IRNN' object has no attribute 'save_for_backward'

thanks

数据集

12

可以给下论文上的两张真实的数据集图片吗？

关于psnr指标

12

您好我用rain100H训练100个epoch后并且测试出来的指标只有PSNR=28.0568，似乎并没有达到论文中29.46这么高