About the normalization for Inception v3 when calculating FID

[AtsukiOsanai]

Thank you for sharing your excellent work!

I'm curious as to why you chose to re-normalize the fake images in this section. Considering that Inception v3 requires input values within [-1, 1], the generated fake images on this line already meet the requirement of inception v3. Therefore, I believe that re-normalizing these images could potentially result in an incorrect FID score if you're using a pre-trained Inception v3 model from torchvision.

From what I understand, re-normalization of the tensor is necessary only when we employ standard normalization as per ImageNet statistics, like in this ImageNet training example.

I hope you will clarify this issue to achieve a fair comparison in image synthesis.

[tobran]

Thanks for your attention to our work. We checked the normalization in our evaluation codes and found that this problem does not exist.

The input of inception v3 is not [-1, 1]. We used the code of DM-GAN to test FID. The image passes through transforms.ToTensor(), and its distribution becomes [0, 1] (https://github.com/MinfengZhu/DM-GAN/blob/d515995b6fade48be9762c7f069c9d889d4d5559/eval/FID/fid_score.py#L222), and then is normalized to the target interval (https://github.com/MinfengZhu/DM-GAN/blob/d515995b6fade48be9762c7f069c9d889d4d5559/eval/FID/fid_score.py#L222).

Our code (

GALIP/code/lib/modules.py

Line 270 in 35d0c2f

fake = norm(fake_imgs)

) converts it from [-1, 1] to [0, 1] (consistent with the distribution after transforms.ToTensor() of DM-GAN), and then normalizes it to the target interval, so the final data distribution input to inception v3 is consistent with DM-GAN.

At the same time, we save the generated images and test them using DM-GAN's evaluation codes. The results are consistent with our code.