Table 4. Effect of stacked layers and order on validation classification accuracy (%). The mean (standard deviation) of the accuracy for three training runs are reported. (GRU, 2x2 receptive field, 32 filters / pass, 64x32 patch size, 25% dropout each layer).

Architecture	Acc. (%)	Architecture	Acc. (%)
Vertical Unidirectional	96.28 (0.02)	Horizontal Unidirectional + Vertical Unidirectional	96.69 (0.02)
2 x Vertical Unidirectional	96.59 (0.02)	Vertical Unidirectional + Horizontal Unidirectional	96.69 (0.02)
3 x Vertical Unidirectional	96.67 (0.01)	Horizontal Bidirectional + Vertical Bidirectional	96.70 (0.01)
		Vertical Bidirectional + Horizontal Bidirectional	96.73 (0.02)