Search is not available for this dataset
audio
audioduration (s)
0.69
35.8
label
class label
3 classes
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
0Chinese
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
1Italian
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
2Japanese
README.md exists but content is empty. Use the Edit dataset card button to edit it.
Downloads last month
267
Edit dataset card

Models trained or fine-tuned on None1145/Vulpisfoglia