search

hila-chefer / Transformer-Explainability

[CVPR 2021] Official PyTorch implementation for Transformer Interpretability Beyond Attention Visualization, a novel method to visualize classifications by Transformer based networks.
MIT License
1.79k stars 241 forks source link

Where can I get the imagenet segmentation dataset? #12

Closed jfloveu closed 3 years ago

jfloveu commented 3 years ago

❔Question Hello, I think it's a very interestring work! But I want to konw where I can get the dataset in "reproducing results on ViT", for example, "gtsegs_ijcv.mat". I hope for your help. Thanks!

hila-chefer commented 3 years ago

Hi @jfloveu thanks for your interest in our work! IMO- we downloaded the segmentation from here: https://github.com/aravindhm/deconvnet_analysis section 7, but I see the link is now corrupt. I’ll check if I can either find you another link or upload it somewhere so you can download it. I’m leaving this issue open, will get back to you with an answer.

jfloveu commented 3 years ago

Thanks for your reply. Of course, other datasets are also OK, as long as they run on your code, which will be more helpful for me to understand how the code works.

hila-chefer commented 3 years ago

Hi @jfloveu,

we are contacting the original authors of the paper we used to get this data. Hopefully, they can share a new and valid link. Since it’s not our data, we cannot upload it to our university servers. I’ll update here once we have a response from them.

hila-chefer commented 3 years ago

@jfloveu an official link: http://calvin-vision.net/bigstuff/proj-imagenet/data/gtsegs_ijcv.mat Taken from: http://calvin-vision.net/bigstuff/proj-imagenet/page/index.html