issues
search
xxxnell
/
how-do-vits-work
(ICLR 2022 Spotlight) Official PyTorch implementation of "How Do Vision Transformers Work?"
https://arxiv.org/abs/2202.06709
Apache License 2.0
798
stars
77
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Can I get a guideline for hessian eigenvalue visualization?
#42
gihyunkim
closed
5 months ago
1
Lesion study
#41
liguopeng0923
closed
6 months ago
1
Understanding loss landscape
#40
SonHyegang
closed
7 months ago
1
relative log magnitude
#39
zhenyuan1234
closed
1 year ago
1
AlterNet on CIFAR10
#38
23Uday
closed
1 year ago
1
pretrained models
#37
chenrunbin123
opened
1 year ago
3
Question about Figure 2(a)
#36
iumyx2612
closed
1 year ago
6
Question about harmonizing Convs with MSAs
#35
iumyx2612
closed
1 year ago
2
Hessian Max eigenvalue spectra 코드 관련 질문드립니다.
#34
Levinna
closed
1 year ago
1
Total parameters in AlterNet
#33
sauravtii
closed
1 year ago
1
pretrained model file is corrupted
#32
youngandbin
closed
1 year ago
1
What factors determine if a model or a layer behaves like a low- or high-pass filter?
#31
waitingcheung
closed
1 year ago
1
Conclusion about long-range dependency seems not true
#30
iumyx2612
closed
1 year ago
6
How would the MSA build-up rules differ for upsampling stages?
#29
waitingcheung
closed
1 year ago
1
ViT vs ResNet: Did you use SAM optimizer?
#28
quannguyen268
closed
1 year ago
1
Findings not compatible with other work?
#27
iumyx2612
closed
1 year ago
4
What exactly makes MSAs data specificity?
#26
iumyx2612
closed
1 year ago
7
Frequency Analysis for MoCo-v3
#25
YuanLiuuuuuu
closed
1 year ago
1
question about detail--drop_pro parameter sd
#24
salt-fisher
closed
1 year ago
1
code about robustness for noise frequency exp (Fig. 2b)
#23
DoyoungYoon
closed
1 year ago
2
Hi
#22
ross-Hr
closed
1 year ago
2
∆ Log amplitude 관련 질문드립니다.
#21
jhcha08
closed
1 year ago
1
Hi，something about object detection...
#20
ross-Hr
closed
1 year ago
1
Update page
#19
hmthanh
closed
1 year ago
0
Image
#18
123456789-qwer
closed
2 years ago
1
Why is feed-forward not present in the paper and the code?
#17
waitingcheung
closed
2 years ago
1
Running time for visualizing the loss landscape
#16
waitingcheung
closed
2 years ago
1
model size
#15
forever10086
closed
2 years ago
5
TensorFlow implementation
#14
andreped
opened
2 years ago
1
How to plot the Trajectories in polar coordinates?
#13
xin-fight
closed
2 years ago
1
How to plot the Hessian max eigenvalue spectra?
#12
Dong1P
opened
2 years ago
7
헤시안 관련해서 질문드립니다.
#11
Yoontae6719
closed
2 years ago
2
Potential mistake in loss landscape visualization.
#10
sjtuytc
closed
2 years ago
3
In convit.py file, where does ConVit come from, really?
#9
dinhanhx
opened
2 years ago
15
Trained models
#8
luluenen
closed
2 years ago
3
what is the attributes in the large-kernel CNN
#7
ccx1997
opened
2 years ago
2
how to compute feature map variances?
#6
LostXine
closed
2 years ago
4
how is robustness calculated?
#5
psteinb
closed
2 years ago
4
Plot for Relative log amplitudes of Fourier transformed feature maps
#4
xingchenzhao
closed
2 years ago
5
Plot for
#3
xingchenzhao
closed
2 years ago
0
Support of the hessian max eigenvalue spectrum code
#2
big-chan
closed
2 years ago
4
Code for Alter-ResNet-50
#1
DarrenIm
closed
2 years ago
5