DNN-5 (Trainable Activations)

Reported on 3 benchmarks across 1 task

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Computer Vision3 results

Image ClassificationonMNIST
Accuracy
97.2
best: 99.87 (Branching/Merging CNN + Homogeneous Vector Capsules)
Image ClassificationonMNIST
Percentage error
2.8
best: 0.13 (Branching/Merging CNN + Homogeneous Vector Capsules)
Image ClassificationonMNIST
Trainable Parameters
575051
best: 1882602 (Neural Architecture Search (NAS)-enabled Convolutional Neural Network (CNN))