hila-chefer Transformer-Explainability issues

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

1.75k stars 232 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

what should i do when trying this visualization method on cifar dataset

#20 caposerenity closed 3 years ago
1
torch._six updates lead to errors in baselines/ViT/layer_helpers.py

#19 philipsgithub closed 3 years ago
2
New post's url about the paper can not be accessed

#18 NickChang97 closed 3 years ago
2
Reproducibility in another text dataset

#17 carlosabcs closed 3 years ago
7
Is it possible to use this tool without [CLS] token?

#16 zeyuyun1 closed 3 years ago
8
Abs in Add Normalization

#15 ChuanyangZheng closed 3 years ago
1
how to integrate CLIP model and find CLIP attention map?

#14 maryamag85 closed 3 years ago
3
Is that possible to calculate the relevance of recurrent neural network (RNN)?

#13 StOnEGiggity closed 3 years ago
2
Where can I get the imagenet segmentation dataset?

#12 jfloveu closed 3 years ago
4
Use my own ViT model

#11 shadowwkl closed 3 years ago
4
missing Normalization in Attention

#10 nkjack closed 3 years ago
2
suport for CLIP

#9 betterze closed 3 years ago
6
Is this possible to be implemented with hybrid model (R50+ViT) ?

#8 gwang-kim closed 2 years ago
9
Multi-card training out of memory

#7 lingorX closed 3 years ago
3
Could you add the baseline Grad-CAM implementation?

#6 ramprs closed 3 years ago
1
if it is possible to support DETR in the future

#5 dc-cheny closed 3 years ago
6
Create LICENSE

#4 hila-chefer closed 3 years ago
0
Question about `no_weight_decay` in the code of ViT

#3 yztongzhan closed 3 years ago
2
missing link to paper

#2 eyaler closed 3 years ago
4
Create LICENSE

#1 hila-chefer closed 3 years ago
0