Table 3.

Feedforward neural network condensation.

Model	Parameters, n	File size (kb)	Inference (seconds per sample)	Training time (seconds;20 epochs)	Test AUROCa (last epoch)
Baseline DNNb	60,929	767	20	3300	0.82
Pruned DNN	27,312	315	10	3310	0.81
Quantized DNN	60,929	64	15	N/Ac	0.82

^aAUROC: area under the receiver operating curve.

^bDNN:

^cN/A: not applicable.