Train Stage 1 Enumeration Error

[pranav-ap]

I am using the CelebVHQ dataset. I ran their preprocessing script. Then extracted the frames into seperate folders - one folder per video.

But when I run the train stage 1 script, I get the following error. Has anyone faced this before?

A truncated version of the training json file looks like this.

{
    "vtznSlhouZc_1": {
        "video_clip_num": 1,
        "clip_data_list": [
            {
                "frame_name_list": [
                    "vtznSlhouZc_1\\1.jpg",
                    "vtznSlhouZc_1\\2.jpg",
                ],
                "frame_path_list": [
                    "F:\\AniPortrait\\data\\praw\\vtznSlhouZc_1\\1.jpg",
                    "F:\\AniPortrait\\data\\praw\\vtznSlhouZc_1\\2.jpg",
                ],
                "lmks_list": [
                    "F:\\AniPortrait\\data\\ppost\\vtznSlhouZc_1\\1_lmks.npy",
                    "F:\\AniPortrait\\data\\ppost\\vtznSlhouZc_1\\2_lmks.npy",
                ]
            }
        ]
    }
}

I tried using next() just to test it out,

So I tried out iter() too,

[Zejun-Yang]

You can try to refer to our dataset processing code (scripts/preprocess_dataset.py).