[Retracted] Improvement of Speech Recognition Technology in Piano Music Scene Based on Deep Learning of Internet of Things

Table 1

Hyperparameters under different data sets.

Hyperparameters	Dataset
	ESC-10	ESC-50	TUT
Dropout ratio	0.18	0.25	0.47
Learning policy	Fixed	Fixed	Stepwise
Momentum	0.025	0.005	0.002
Weight decay	0.89	0.94	0.88
Learning rate	Adam	SGD	SGD
Mini-batch	0.05	0.01	0.02
Optimizer type	31	117	31
Epochs	49	78	149