Short Utterance and Cross Lingual Speaker Verification
The publisher made the trial set
veri_test.txt is the baseline from Voxceleb2. It is downloadable from Voxceleb website
short.txt is from the google speech command data
cross_lingual.txt is the partial data from Common Voice consists of English, German, French, Spanish, Japanese, Russian, Dutch
| Voxceleb-O | Google Speech Command | Common Voice |
|---|---|---|
| 4708 files | 2276 files | 2226 files |
| 40 speakers | 282 speakers | 788 speakers |
| 4708 sentences | 35 tokens | 83 tokens |
