Datasets:
Modalities:
Audio
Sub-tasks:
speaker-identification
Languages:
Japanese
Size:
100K<n<1M
Tags:
speech-recognition
License:
-
27.8 kB
-
14.1 kB
-
24.9 kB
-
26.8 kB
-
10.3 kB
-
26.8 kB
-
18.6 kB
-
21.5 kB
-
11.8 kB
-
10.9 kB
-
29.7 kB
-
22.6 kB
-
24.7 kB
-
24.1 kB
-
25.8 kB
-
22.5 kB
-
12.6 kB
-
25 kB
-
27.5 kB
-
20.7 kB
-
19 kB
-
22.7 kB
-
23.6 kB
-
30.4 kB
-
25.5 kB
-
28.2 kB
-
23.1 kB
-
23.2 kB
-
25.5 kB
-
26 kB
-
22.7 kB
-
26.8 kB
-
23.1 kB
-
21.8 kB
-
24 kB
-
23 kB
-
25.3 kB
-
22.8 kB
-
23.8 kB
-
22.2 kB
-
19.6 kB
-
21.1 kB
-
24.2 kB
-
18.8 kB
-
19.3 kB
-
11.6 kB
-
21.2 kB
-
9.67 kB
-
24.1 kB
-
19.8 kB