Discriminator output size

[ChristianEschen]

Hi

I have found some difficulties understanding the receptive field in the discriminator. The discriminator takes a 256x256 input and discriminates 30x30 patches with receptive field equal 70x70. I see you use the following equation to calculate the receptive field size. inputsize=(outputsize-1)*stride+kernelsize (1)

I guess this equation is derived from outputsize=(inputsize-kernelsize+paddingwidth+paddingheight)/stride + 1 (2) . So how do you pad the images. From the code it looks like paddingwidth=1 and paddingheight=1. However, this does not make sense to me with respect to equation (1). I hope you can help me out here.

Best regards

[litingfeng]

I have a concern too. Here should we understand RECEPTIVE FIELD as the input size, rather than the real 'receptive field'? I think the default padding=1 in this case. Here is the link to calculate input/output size, and receptive field.