may you share the CSL dataset

Question

may you share the CSL dataset

Closed this issue 2 years ago · 7 comments

as title, maybe Baidu Netdisk link， thanks a lot.

Answer 1 · 2023-02-26T07:31:36.000Z

Sorry, but according to the data release agreement that i signed, i can't share the data by my own. You could follow this web to request the data.

Answer 2 · 2023-03-05T03:36:31.000Z

I got the datasets, but i have some trouble in data_preprocess-CSL.py . how can i get the tran.txt, dev.txt, and voc.txt.
And thanks a lot,

Answer 3 · 2023-03-05T07:53:46.000Z

train.txt, dev.txt and voc.txt are here.

Answer 4 · 2023-03-05T08:09:49.000Z

thanks for your relply,I mean that i get the CSL_Continuous data just contains videos and a corpus.txt. But in data_preprocess-CSL.py your datas all endwith .jpg. If you do some pre-process, would you mind to share it, and thanks again.

Answer 5 · 2023-03-05T08:11:37.000Z

supplementary

Answer 6 · 2023-03-05T08:21:00.000Z

The data i got is consisted of .jpg files. I also notice that the video names of yours are not the same with mine. I suspect that they have changed over time, but with the same contents. Maybe you should rewrite the data-preprocess.py according to the label file you get.

Answer 7 · 2023-03-05T08:36:02.000Z

Thanks for sharing ! Expect the following works.In addition, if now you shares jpgs data is not breach of release agreement when i have access to got the video data.
email : creepui@gmail.com