Continual Learning on visual domain decathlon (10 tasks)

Metric: decathlon discipline (Score) (higher is better)

LeaderboardDataset

Loading chart...

Results

Sort:

#	Model↕	decathlon discipline (Score)▼	Augmentations	Paper	Date↕	Code
1	NetTailor	3744	No	NetTailor: Tuning the Architecture, Not Just the...	2019-06-29	Code
2	Depthwise Soft Sharing	3507	No	Depthwise Convolution is All You Need for Learni...	2019-02-03	Code
3	Parallel Res. adapt.	3412	No	Efficient parametrization of multi-domain deep n...	2018-03-27	Code
4	Depthwise Sharing	3234	No	Depthwise Convolution is All You Need for Learni...	2019-02-03	Code
5	Series Res. adapt.	3159	No	Efficient parametrization of multi-domain deep n...	2018-03-27	Code
6	Res. adapt. (large)	3131	No	Learning multiple visual domains with residual a...	2017-05-22	Code
7	DAN	2851	No	Incremental Learning Through Deep Adaptation	2017-05-11	-
8	Piggyback	2838	No	Piggyback: Adapting a Single Network to Multiple...	2018-01-19	Code
9	Res. adapt. finetune all	2643	No	Learning multiple visual domains with residual a...	2017-05-22	Code
10	Res. adapt. decay	2621	No	Learning multiple visual domains with residual a...	2017-05-22	Code
11	LwF	2515	No	Learning without Forgetting	2016-06-29	Code
12	Res. adapt. dom-pred	2503	No	Learning multiple visual domains with residual a...	2017-05-22	Code
13	Res. adapt.	2118	No	Learning multiple visual domains with residual a...	2017-05-22	Code
14	BN adapt.	1363	No	Universal representations:The missing link betwe...	2017-01-25	-

#1NetTailorSOTA
3744
decathlon discipline (Score)· 2019-06-29
NetTailor: Tuning the Architecture, Not Just the Weights Code
#2Depthwise Soft SharingSOTA
3507
decathlon discipline (Score)· 2019-02-03
Depthwise Convolution is All You Need for Learning Multiple Visual Domains Code
#3Parallel Res. adapt.SOTA
3412
decathlon discipline (Score)· 2018-03-27
Efficient parametrization of multi-domain deep neural networks Code
#4Depthwise Sharing
3234
decathlon discipline (Score)· 2019-02-03
Depthwise Convolution is All You Need for Learning Multiple Visual Domains Code
#5Series Res. adapt.
3159
decathlon discipline (Score)· 2018-03-27
Efficient parametrization of multi-domain deep neural networks Code
#6Res. adapt. (large)SOTA
3131
decathlon discipline (Score)· 2017-05-22
Learning multiple visual domains with residual adapters Code
#7DANSOTA
2851
decathlon discipline (Score)· 2017-05-11
Incremental Learning Through Deep Adaptation
#8Piggyback
2838
decathlon discipline (Score)· 2018-01-19
Piggyback: Adapting a Single Network to Multiple Tasks by Learning to Mask Weights Code
#9Res. adapt. finetune all
2643
decathlon discipline (Score)· 2017-05-22
Learning multiple visual domains with residual adapters Code
#10Res. adapt. decay
2621
decathlon discipline (Score)· 2017-05-22
Learning multiple visual domains with residual adapters Code
#11LwFSOTA
2515
decathlon discipline (Score)· 2016-06-29
Learning without Forgetting Code
#12Res. adapt. dom-pred
2503
decathlon discipline (Score)· 2017-05-22
Learning multiple visual domains with residual adapters Code
#13Res. adapt.
2118
decathlon discipline (Score)· 2017-05-22
Learning multiple visual domains with residual adapters Code
#14BN adapt.
1363
decathlon discipline (Score)· 2017-01-25
Universal representations:The missing link between faces, text, planktons, and cat breeds