embedder.pt with new dataset

[Devid29]

Hi, if in case I wanted to use another dataset of audio files for the training and the test (not the one used here) the embedder.pt that I have to insert when I run "trainer.py" as I can generate it or which one I have to use ? Thank you

[seungwonpark]

Generally, you can simply use the provided embedder.pt for all cases. Since the speaker vector extractor was trained on VoxCeleb2 dataset, it works well with various language/environments. Seung-won Park Undergraduate Student Dept. of Physics and Astronomy & Computer Science Engineering Seoul National University yyyyy@snu.ac.kr | http://swpark.me

…

________________________________ From: Devid29 <notifications@github.com> Sent: Thursday, September 12, 2019 12:36:46 AM To: mindslab-ai/voicefilter <voicefilter@noreply.github.com> Cc: Subscribed <subscribed@noreply.github.com> Subject: [mindslab-ai/voicefilter] embeddet.pt with new dataset (#19) Hi, if in case I wanted to use another dataset of audio files for the training and the test (not the one used here) the embedder.pt that I have to insert when I run "trainer.py" as I can generate it or which one I have to use ? Thank you — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#19>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ADGGC5R3YU7Y3PUFZ5GS4NLQJEGA5ANCNFSM4IVVT3WA>.

[Devid29]

Ok thanks because the dataset I am using is in Italian and I didn't know if this could create problems. Thank you

[seungwonpark]

Yes, there were no significant problem when I was trying to use it with Korean speech. Good luck on your testing with Italian speech!

Best, Seungwon

[Devid29]

Thank you!