sberdevices / golos

208 stars 13 forks source link

Speaker labels #5

Closed maksim-pankov closed 3 years ago

maksim-pankov commented 3 years ago

Is it possible to add speaker Ids (depersonalized) of any kind to dataset jsonl files? I need them to distinguish utterances from different speakers to train the voice embedding model (for speaker identification and speaker diarization tasks).

AlexDenis commented 3 years ago

Sadly we don't have such information for this dataset