Make-A-Story icon indicating copy to clipboard operation
Make-A-Story copied to clipboard

Published 20 hours ago • verified publisher icon ubc-vision

Extended dataset

Open xiaoqian-shen opened this issue 1 year ago • 5 comments

Great work! I am curious about whether you will release the extended dataset with referential text.

xiaoqian-shen avatar Jul 06 '23 11:07 xiaoqian-shen

I would also appreciate the release of this dataset

snarkyidiot avatar Jul 11 '23 07:07 snarkyidiot

Btw, could you provide the pretrained character classifier for convenient evaluation?

xiaoqian-shen avatar Jul 11 '23 11:07 xiaoqian-shen

We re-used existing datasets. In the dataloader, we modify the ref texts. See ldm/data/flintstones_data.py, ldm/data/mugen_data.py and ldm/data/proro_data.py to generate ref text.

trahman8 avatar Jul 25 '23 23:07 trahman8

Thanks for your reply. This issue then could be closed.

xiaoqian-shen avatar Jul 26 '23 16:07 xiaoqian-shen

Hi! I am trying to evaluate the accuracy of generated stories. Could you provide the pretrained weight of character and background classifier for convenient evaluation?

xiaoqian-shen avatar Sep 11 '23 12:09 xiaoqian-shen