Evaluating the Performance of TAAF for image classification models

Benchmark Model Rank Results
image-classification-on-cifar-10The Analog Activation Function#215Percentage correct: 82.06Parameters: 545100Top 1 Accuracy: 82.06
image-classification-on-mnistTAAF-CNN#24Percentage error: 0.48%Accuracy: 99.52%Trainable Parameters: 421642