BFFHQ
Benchmarks
TaskDatasetSOTA resultTrendResults
BFFHQ (test)
71.33Accuracy (BC)
18
BFFHQ bias-conflicting (test)
76.8Accuracy
17
BFFHQ
8.53Relative Accuracy Improvement
14
BFFHQ 0.5% bias ratio unbiased (test)
65.8Accuracy (Minority)
11
BFFHQ (test)
0.717Accuracy @ Thresh 0.5
11
BFFHQ 5.0% bias-conflict ratio
82.73BC Score
9
BFFHQ 1.0% bias-conflict ratio
51.48Bias Conflict Rate
9
BFFHQ unbiased (test)
87.57Accuracy
6
BFFHQ ratio 1.0 (test)
75.7Accuracy
6
bFFHQ bias-conflicting samples (test)
62.8Accuracy
6
BFFHQ
63Accuracy
5
bFFHQ (Bias-guiding)
99.13Accuracy
4
BFFHQ
64EOD
2
BFFHQ (test)
5.52Mean Relative Improvement (Avg Acc)
2