search

MCG-NJU / TRACE

[ICCV 2021] Target Adaptive Context Aggregation for Video Scene Graph Generation
Other
58 stars 5 forks source link

Link for downloading features for VIDVRD dataset cannot be accessed #5

Closed Harryqu123 closed 3 years ago

Harryqu123 commented 3 years ago

I find this an interesting work and I would like to reimplement your work following the code. However, I find that I cannot access the link you provided in readme for features for VIDVRD dataset(https://internal-api-drive-stream.larksuite.com/space/api/box/stream/download/all/boxusS8Z0kwEizoPPh5h7vx7Usf/). Can I know is there any other way for me to access it? Thanks in advance

Harryqu123 commented 3 years ago

I think for a reason that I am not sure I can access and download the features for VIDVRD dataset now. However, when I try to dump frames following the readme, I find that both the val_fname_list.json and the train_fname_list.json is neither in the github repo and nor in the files that are required to be downloaded. Hence, can I ask How should I get such two files?

tyshiwo1 commented 3 years ago

sorry for my mistake. You may 'Change the format of annotations for AG and VidVRD' first and then dump frames. I've fixed the mistake in readme.

tyshiwo1 commented 3 years ago

As for the download link, I had checked it and found it worked before, but it seems not work now. I will upload the features recently.

tyshiwo1 commented 3 years ago

I've added my download link

Harryqu123 commented 3 years ago

Sorry for the late reply. Thanks for your reply very much, I think they solve my issue already and I would close the comment. Many thanks