Fashion MNIST
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
Fashion-MNIST 4x2000 (test)
87.12Accuracy
11
Fashion-MNIST 2% (test)
86.56Accuracy
11
Fashion-MNIST 2% subset
0.942Accuracy
11
Fashion-MNIST (test)
84.13Accuracy
11
Fashion-MNIST (retain)
95.66Acc
11
Fashion-MNIST (Entire data stream)
71.26Accuracy
10
Fashion-MNIST
84.98A_theta Robustness Score
10
Fashion-MNIST MINILENET (test)
99.16Pretrain Accuracy
10
Fashion-MNIST MINILENET
49.07Pretrain Accuracy
10
Fashion MNIST
0.215Fidelity (Uniform)
10
Fashion-MNIST 1k
58.4ACC
10
Fashion MNIST
54.37Intrinsic Dimension
9
Fashion-MNIST non-IID alpha=0.5
90.37CBA MTA
9
Fashion-MNIST
1.764Runtime (s)
9
Fashion-MNIST unseen classes 5-9
84.7Accuracy
9
Fashion-MNIST Dir(0.01), 50 clients, 20% participation
82.95Acc (SQ 4bit)
9
Fashion-MNIST Dir(0.01), 10 clients, full participation
84.17Accuracy (SQ 4bit)
9
Fashion-MNIST Path(1), 10 clients, full participation
84.04Accuracy (4bit Quantization)
9
Fashion-MNIST gamma=3 (Pathological)
85.31Accuracy
9
Fashion-MNIST gamma=6 (Pathological)
88.05Accuracy
9
Fashion-MNIST beta=0.1 (Dirichlet)
84.64Accuracy
9
Fashion-MNIST Dirichlet beta=0.6
88.46Accuracy
9
Fashion-MNIST dynamically binarized (test)
-234.19Test Log-Likelihood Bound
9
Fashion-MNIST dynamically binarized (train)
-234.11ELBO (Train)
9
Fashion-MNIST-1k
0.633NMI
9