Table 4.

Average processing time (5 runs), for S3 dataset single speaker model, varying number of LSTM layers.

LSTM Layers	Mean Training Time (Minutes)	IQR
3-layers	41.75	14.62
4-layers	62.95	18.32
5-layers	78.40	17.25
6-layers	88.07	20.77
7-layers	103.40	11.94
8-layers	139.12	45.23
9-layers	138.80	61.53