Naiyuan Liu

About DF-VAE

5

Thank you for your work! Do you intend to release the relevant code and training scripts about the DF-VAE?

Thanks for sharing your work! I have followed the gudience of Traning V2 to download the IJBC dataset.But I found there are some many subdir in the base root, including...

About the segmentation model

1

Thanks for sharing your work. where can I get model: "celeba_unet_256_1_2_segmentation_v2.pth"

Why MLP uses fully-connected blocks instead of convolution blocks？ What are the advantages of fully-connected blocks over convolution blocks？Could you elaborate on that.

About the clean version of FF++ dataset you used in Quantitative Comparison.

1

Thanks for sharing your work. It is really marvelous! As you mentioned in the "Experiments on FaceForensics++" section: "However, as FaceForensics++ is not designed for face recognition, some videos display...

About non-id space training.

1

Thanks for sharing your work.Could you tell me how you train the non-id speace and what kind of loss did you use.Cause I find that there are seems not loss...

下载速度慢，开了svip也只有3m每秒

1

我已经开了svip，同样一个电影命令行下载速度是3M，客户端下载速度是11M，如下图所示，请问是我有什么没有设置好吗

如何直接下载别人分享的链接

1

谢谢作者开源这个仓库！我看了一下readme，只看到下载自己网盘下的内容，没法直接通过网盘分享的链接下载。如果要通过网盘分享的链接直接下载，是需要先转存到自己的网盘目录下，然后再下载吗，如何确认转存之后的文件名或者文件夹名呢

Confusion about why only using the first frame of the video to crop the face region for the stablized image generation?

1

Thanks for sharing your amazing work. In the first step, you crop the face of each frame following the ffhq style. But I notice that you use the coordinates of...