Application of Transfer Learning Algorithm and Real Time Speech Detection in Music Education Platform

Table 2

Speech recognition accuracy after applying the fine-tuning of frozen layer.

Research objects	Class hours	Percentage of frozen fine adjustment layer
		10	30	50	70	90
Female/Male	25	64.29	61.29	61.52	54.30	23.60
	50	63.32	62.94	59.31	54.10	29.11
	100	63.89	63.62	61.36	53.29	34.30
	200	62.22	61.42	57.10	53.80	35.64
Female sex	25	76.25	77.20	74.61	65.14	35.58
	50	71.96	76.56	72.58	65.61	41.20
	100	76.62	76.36	74.25	65.72	39.76
	200	74.90	75.28	71.58	66.40	46.50
Male	25	73.10	78.66	76.79	71.24	25.10
	50	77.32	76.74	74.88	65.70	39.78
	100	77.80	76.87	74.71	66.68	43.86
	200	75.32	76.02	71.66	65.56	39.50