mgrankin over9000 issues

mgrankin / over9000

Over9000 optimizer

Apache License 2.0

425 stars 57 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Please update README with latest optimezers like MADGRAD and AdaBelief

#25 shkarupa-alex closed 3 years ago
2
Performance mismatch between notebook and readme

#24 juntang-zhuang closed 3 years ago
2
If I use RangerLars should I use any additional lr_scheduler

#23 tugot17 closed 4 years ago
1
Recommendation on optimal LR

#22 IamGianluca closed 4 years ago
3
RangerLars

#21 sky186 closed 4 years ago
2
Update benchmarcks now that Ranger supports gradient centralization

#20 LifeIsStrange closed 4 years ago
1
Port to fastai v2?

#19 oguiza closed 4 years ago
6
How about RAdam + LAMB + Lookahead?

#18 Whu-wxy closed 4 years ago
2
Create LICENSE

#17 mgrankin closed 4 years ago
0
What do you think about weight decay?

#16 hadaev8 closed 4 years ago
1
Citing the repository

#15 kakumarabhishek closed 4 years ago
2
What is the meaning of Imagenette and Imagewoof ?

#14 fyting closed 5 years ago
2
LR Scheduler

#13 kakumarabhishek closed 5 years ago
2
question about LAMB implementation.

#12 fastalgo closed 4 years ago
7
Merge them all?

#11 LifeIsStrange closed 5 years ago
1
Applying fix for different param_groups for RAdam

#10 redknightlois closed 5 years ago
2
Fix bug in RAdams buffer implementation

#9 sholderbach closed 5 years ago
2
Adding state_dict to LookAhead optimizer

#8 LeviViana closed 5 years ago
2
Proper implementation of Ralamb

#7 redknightlois closed 5 years ago
0
Lookahead has no attribute 'state'

#6 staylor-ds closed 5 years ago
2
Testing that uses LAMB only

#5 Tony-Y closed 5 years ago
2
AdaBound

#4 r1ckya closed 4 years ago
6
Failure to execute on Windows.

#3 redknightlois closed 4 years ago
12
Updated Ralamb Source

#2 redknightlois closed 5 years ago
2
How to use over9000 in common pytorch code?

#1 askerlee closed 5 years ago
10