Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

ObjectNet

Benchmarks

Task NameDataset NameSOTA ResultTrend
Image ClassificationObjectNet
Accuracy91.6
251
Class-Incremental LearningObjectNet
Average Accuracy83.8
60
Image ClassificationObjectNet (test)
R@152
43
Zero-shot Image ClassificationObjectNet
Zero-shot Accuracy29.16
18
Machine UnlearningObjectNet
Accuracy29.16
18
Image ClassificationObjectNet (val)
Top-1 Acc87.6
12
Selective classificationObjectNet (ON)
AURC0.333
11
Image ClassificationObjectNet-9 distribution-shifted (test)
Test Accuracy31.38
10
General Object RecognitionObjectNet 313 (test)
Exact Match (EM)24.4
9
Uncertainty Quantification for Object Classification in Unusual PosesObjectNet (test)
Rejection Accuracy @ 90%68.7
9
Image ClassificationObjectNet (OOD)
Top-1 Acc39.9
9
Class-Incremental LearningObjectNet B100 Inc20
Average Performance54.15
7
Class-Incremental LearningObjectNet B0 Inc20
Average Performance61.02
7
Class Incremental LearningObjectNet Inc-10 B-0 (test)
Favg11.2
7
Class Incremental LearningObjectNet Inc-5 B-0 (test)
Favg11.55
7
Class-Incremental LearningObjectNet 20 steps (incremental)
Average Accuracy68.12
7
Image ClassificationObjectNet 1.0 (test)
Top-1 Accuracy72.3
7
OOD DetectionObjectNet-104 (test)
AUC99.3
6
OOD DetectionObjectNet-54 (test)
AUC0.986
6
Untargeted Score-based Black-box AttackObjectNet
ASR100
6
Targeted Score-based Black-box AttackObjectNet
ASR57.5
6
Class Incremental LearningObjectNet Inc10 (test)
Average Accuracy71
6
Image ClassificationObjectNet 113 shared classes
Accuracy69.2
5
Image ClassificationObjectNet latest (test)
Top-1 Acc26.6
5
Class Incremental LearningObjectNet B-0
Last Accuracy (Alast)67.53
4
Showing 25 of 30 rows