issues
search
sIncerass
/
powernorm
[ICML 2020] code for "PowerNorm: Rethinking Batch Normalization in Transformers" https://arxiv.org/abs/2003.07845
GNU General Public License v3.0
119
stars
17
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
A few questions regarding fairseq/modules/norms/mask_powernorm.py
#15
congwang093
opened
1 year ago
18
Comparisons with RMSNorm？
#14
xiaoxin83121
closed
1 year ago
0
Question regarding the batch norm vs masked batch norm
#13
enhuiz
opened
1 year ago
0
a question about the image of layer normalization in README.md
#12
erjiaxiao
closed
2 years ago
1
Is MaskPowerNorm the powernorm proposed by the paper?
#11
SunTongtongtong
closed
2 years ago
1
PowerNorm link broken
#10
hello-wzy
closed
3 years ago
1
Does PowerNorm still work for NMT task after removing the GroupScaling layer?
#9
CheerM
closed
3 years ago
4
Why use group scaling?
#8
htwang14
closed
3 years ago
2
Cannot reproduce the results on IWSLT14.
#7
ghost
closed
3 years ago
1
Feature request: improved documentation
#6
Guitaricet
opened
3 years ago
3
ImportError: cannot import name 'libbleu' from 'fairseq'
#5
XuMengyaAmy
closed
3 years ago
1
The broken affine parameter and the redundancy codes
#4
grassking100
closed
3 years ago
1
Different backward implementation from the content written in paper
#3
ojm9898
closed
3 years ago
4
Gradient overflow (NaN problem)
#2
CODEJIN
closed
3 years ago
2
Language Modelling code?
#1
arvieFrydenlund
closed
3 years ago
1