issues
search
facebookresearch
/
dadaptation
D-Adaptation for SGD, Adam and AdaGrad
MIT License
499
stars
19
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Running issues of d_hat < 0
#42
zhujiem
closed
6 months ago
3
TypeError: DAdaptLion.__init__() got an unexpected keyword argument 'decouple'
#41
DarkAlchy
closed
7 months ago
2
Support layer-wise scaling
#40
adefazio
closed
9 months ago
1
Small bug fix: change default eps in DAdaptAdaGrad to avoid value error
#39
brunzema
closed
10 months ago
1
dadapt_adam.py Shouldn't it be Ak+1 here?
#38
lu-minous
closed
11 months ago
0
No easy way to read the `dlr` during the training
#37
mitalbert
closed
11 months ago
3
Higher momery usage when using DAdaptAdam
#36
mrembalski
closed
1 year ago
1
Dadapt Lion feedback
#35
Goldenkoron
closed
1 year ago
6
doc and print fixes
#34
adefazio
closed
1 year ago
0
Small mistakes in DAdaptLion class
#33
pamparamm
closed
1 year ago
1
Fix for d_hat logging in lion #30
#32
adefazio
closed
1 year ago
0
Correct default eps to 0.0
#31
kozistr
closed
1 year ago
1
Exception in Lion optimizer
#30
tjennings
closed
1 year ago
0
Learning rates
#29
RealBeyondMaster
closed
1 year ago
2
Remove betas from DAdaptSGD documentation
#28
rockerBOO
closed
1 year ago
2
SGD default momentum is 0
#27
rockerBOO
closed
1 year ago
3
No momentum for DAdaptAdam
#26
rockerBOO
closed
1 year ago
3
Steps per Epoc in Sample images?
#25
DarksealStudios
closed
1 year ago
5
Support for Dylora
#24
sdbds
closed
1 year ago
15
V3
#23
adefazio
closed
1 year ago
0
Added Lion
#22
adefazio
closed
1 year ago
0
Fix `exp_avg_diff` and `exp_avg_sq` assignment in D-Adaptation Adan
#21
kozistr
closed
1 year ago
1
Add Lion with d-adaptation
#20
nviwch
closed
1 year ago
7
Fully-sharded data parallel support
#19
adefazio
closed
1 year ago
0
First inequality on page 3 derivation?
#18
mega-optimus
closed
1 year ago
2
Any chance TF version will be realeased?
#17
georgeYanch
closed
1 year ago
3
D-Adaptation Adan
#16
qwopqwop200
closed
1 year ago
1
Complex number not supported for max() operation
#15
jinyi12
closed
1 year ago
3
[Suggestions]D-Adaptation for Adan(Adaptive Nesterov Momentum Algorithm)
#14
qwopqwop200
closed
1 year ago
1
Bugfix/call with no grads
#13
ariaameri
closed
1 year ago
4
Complex number support
#12
adefazio
closed
1 year ago
1
Any tips for training on my own dataset / fixing scale from COLMAP?
#11
nlml
closed
1 year ago
2
'<' not supported between instances of 'float' and 'complex'
#10
jinyi12
closed
1 year ago
4
fix improper capitalization
#9
ksadov
closed
1 year ago
1
Import Error
#8
stefanwezel
closed
1 year ago
3
Doc and Naming
#7
adefazio
closed
1 year ago
0
Float 16?
#6
TKassis
closed
1 year ago
1
Correctly spell optimizer class names
#5
HelgeS
closed
1 year ago
1
ZeroDivisionError: float division by zero
#4
egorsmkv
closed
1 year ago
5
Pip restructuring
#3
adefazio
closed
1 year ago
0
Pip package or official PyTorch integration?
#2
TKassis
closed
1 year ago
2
missing gsq_weighted in dadapt_sgd
#1
mortadhaaj
closed
1 year ago
1