search

winddori2002 / TriAAN-VC

TriAAN-VC: Triple Adaptive Attention Normalization for Any-to-Any Voice Conversion
MIT License
129 stars 12 forks source link

a question about the evaluation results #23

Open kay9012 opened 1 month ago

kay9012 commented 1 month ago

Is the evaluation data in the thesis table in the article obtained with main.py test after training, or is it obtained by evaluating the results using the conevert? And how to distinguish between s2s and u2u?

winddori2002 commented 1 month ago

Hi,

The results are obtained by running main.py test after training.

You can refer to generate_eval_pair.py to distinguish s2s and u2u. It generates metafiles for testing s2s and u2u. It splits speakers for s2s and u2u for the VCTK dataset.

Thanks.

kay9012 commented 1 month ago

Hi,

The results are obtained by running main.py test after training.

You can refer to generate_eval_pair.py to distinguish s2s and u2u. It generates metafiles for testing s2s and u2u. It splits speakers for s2s and u2u for the VCTK dataset.

Thanks.

Thank you for your responding! ! And how about the Speaker Verification? I use the Resemblyzer/demo05_fake_speech_detection.py to get it. Is it correct?

winddori2002 commented 1 month ago

I think it uses a similar process. You can check ./src/metric.py to get scores related to speaker verification.

Thanks.