GMvandeVen continual-learning issues

GMvandeVen / continual-learning

PyTorch implementation of various methods for continual learning (XdG, EWC, SI, LwF, FROMP, DGR, BI-R, ER, A-GEM, iCaRL, Generative Classifier) in three different scenarios.

MIT License

1.53k stars 310 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Results for None ("lower target")

#31 youtao1 opened 3 months ago
0
About printing results of experimental output

#30 youtao1 closed 3 months ago
0
Whether context identity must be inferred in case of domain increment？

#29 youtao1 closed 5 months ago
1
0 accuracy values for task-free setting

#28 hiteshvaidya closed 8 months ago
9
Single head or multihead task incremental

#27 COD1995 closed 10 months ago
1
Task-IL evaluation

#26 Young-Jo-Choi closed 11 months ago
2
Joint training results different for different types of incremental learning?

#25 toshi2k2 closed 11 months ago
3
Add support for FL

#24 malikDaCoda closed 1 year ago
0
How to create Resnet34

#23 Peeaak closed 1 year ago
2
about kafc fisher infromation matrix

#22 fmaaf opened 1 year ago
1
Reproducing BI+SI method

#21 valeriya-khan closed 1 year ago
9
Link error

#20 MingkangXiong closed 1 year ago
2
Suspicious Precision

#19 tarasrumezhak closed 2 years ago
3
one little confusion about the loss_fn_kd function

#18 libo-huang closed 2 years ago
1
Lower/Upper Bound Experiments

#17 kaustubholpadkar closed 3 years ago
2
why batch_size has to be 1 when update fisher?

#16 yuchenlin closed 2 years ago
1
Wrong dataset?

#15 GuillaumeLam closed 3 years ago
2
Grad in SI

#14 ssydasheng opened 3 years ago
4
Just a request

#13 jaytimbadia closed 3 years ago
0
Datasets more complicated than MNIST

#12 xu-ji closed 4 years ago
1
Empirical Fisher Estimation

#11 tangbinh closed 4 years ago
3
Performance

#10 Johswald closed 4 years ago
7
Question about Online EWC

#9 DavideA closed 5 years ago
2
How to apply iCaRL on task and domain scenario?

#8 hsqzzpf closed 5 years ago
2
Why uses the prediction of all active classes?

#7 suanrong closed 5 years ago
4
average EWC pickle

#6 Johswald closed 5 years ago
6
Knowledge Distillation Loss

#5 Johswald closed 5 years ago
2
permutedMNIST accs

#4 Johswald closed 5 years ago
2
Final results of class incremental learning

#3 Harry-Up closed 5 years ago
4
iCaRL's accuracy different from 'Three scenarios for continual learning'

#2 JosephKJ closed 5 years ago
2
Thank you for your wonderful package!

#1 JosephKJ closed 5 years ago
2