Ask about feature maps modulation

[hueledao]

Hello Thank you for your work.

I saw in SFT or SPADE papers, they modified the feature map by f_hat = f . alpha + beta

Could you explain why in your paper and your code, the modified features is computed by f_hat = f . (1+alpha) + beta

Any reasons and advantages for computing like that? Thank you again.

[IceClear]

I do not think there is an obvious difference between them.