Sound Classification Based on Multihead Attention and Support Vector Machine

Table 1

Dataset description.

	UrbanSound8K		GTZAN		IEMOCAP
	Genre	Samples (#)	Genre	Samples (#)	Genre	Samples (#)
0	Air condition (AI)	1000	Classic	100	Happy	1636
1	Car horn (CA)	429	Jazz	100	Sad	1084
2	Children playing (CH)	1000	Blues	100	Angry	1103
3	Dog dark (DO)	1000	Metal	100	Neutral	1708
4	Drilling (DR)	1000	Pop	100
5	Engine idling (EN)	1000	Rock	100
6	Gun shot (GU)	374	Country	100
7	Jackhammer (JA)	1000	Disco	100
8	Siren (SI)	929	Hip-hop	100
9	Street music (ST)	1000	Reggae	100
	Total	8732		1000		5531