Home
Ontology
Dataset
Download
About
Dataset
Singing
Musical sounds produced with the human voice, including sustained tonality, rhythm, and a variety of other vocal techniques.
Wikipedia
View full entry in ontology
Class breakdown
Dataset
Number of videos
Duration hours
Overall
42,493
118.0
Evaluation
362
1.0
Balanced train
517
1.4
Unbalanced train
41,614
115.6
Learn more about the dataset parts.
✓
We estimate this class has high quality
In a random sample of videos for this class, we found 9 / 9 (100%) were accurate. Note that quality in the unbalanced training set may be significantly lower.
Learn more
.
Evaluation videos for Singing
362
videos
1.0
hours
View all 362 videos
Balanced train videos for Singing
517
videos
1.4
hours
View all 517 videos
Unbalanced train videos for Singing
41,614
videos
115.6
hours
View all 41,614 videos