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

BloodMNIST

Benchmarks

Task NameDataset NameSOTA ResultTrend
8-class classificationBloodMNIST
Accuracy99
33
Medical Image ClassificationBloodMNIST
AUC100
18
Image ClassificationBloodMNIST balanced (test)
Balanced Accuracy98.5
16
Federated Image ClassificationBloodMNIST (test)
Average Global Accuracy64.5
12
Image ClassificationBloodMNIST v2 (test)
AUC99.8
12
ClassificationBloodMNIST v1 (test)
Accuracy96
12
ClassificationBloodMNIST (test)
Coverage91.2
11
Medical Image ClassificationBloodMNIST
Accuracy70.15
10
Image ClassificationBloodMNIST 2-way sharded (test)
Test Accuracy91
9
Image Classification CalibrationBloodMNIST
NLL0.3119
9
Out-of-Distribution DetectionBloodMNIST
ID Accuracy90.21
8
Open-Set RecognitionBloodMNIST t=5 (test)
Accuracy98.5
8
Model Editingbloodmnist
ΔACC (pp)3.27
6
Image ClassificationBloodMNIST C
Accuracy77
6
Image ClassificationBloodMNIST
Top-1 Accuracy96.23
4
Image ClassificationBloodMNIST MedMNIST (test)
Accuracy95.01
2
Multi-Instance Learning ClassificationBloodMNIST v2 (test)
Metric-
0
Showing 17 of 17 rows