prepare KTH dataset

[zzzzzx-1115]

Hello, I am trying to train SimVP in KTH, but I find the tool for KTH seems not enough.

After unzipping those .zip files, I got a lot of .avi videos. However, according to the dataloader script, I think under directories like 'boxing', there should be some images of shape (1000, 1000), but I only got .avi consisting of frames of shape (120, 160).

Could you please tell me if there are any steps I forget to take to perform training in KTH dataset? Thank you!

[Lupin1998]

Hi, @zzzzzx-1115, thanks for using OpenSTL. Sorry for the bug in the KTH dataset preparation in download_kth.sh, the data should be in image format. You can download the KTH dataset from tsinghua cloud, and all frames under the .avi folder should be $160\times 120$ resolutions. We resize the frame to $128\times 128$ for the KTH experiments. We will fix download_kth.sh in the next commit.

[Lupin1998]

Hello, we have fixed the preparation of the KTH dataset and also uploaded kth_action.zip to Baidu Cloud. Please download the image version of the KTH action, and feel free to raise questions if more issues occur.

[zzzzzx-1115]

Thanks for your response! I will try later.