search

modelscope / 3D-Speaker

A Repository for Single- and Multi-modal Speaker Verification, Speaker Recognition and Speaker Diarization
Apache License 2.0
1.07k stars 93 forks source link

how to extract embeddings by ecapa or resnet #66

Closed JuneAndJuly closed 6 months ago

JuneAndJuly commented 6 months ago

The old model extracting speaker features is no longer supported

yfchenlucky commented 6 months ago

Could you please provide a detailed description of your issue? Is it an error with the code or with the model?

JuneAndJuly commented 6 months ago

Could you please provide a detailed description of your issue? Is it an error with the code or with the model?

image image I extracted the speaker vector based on this, but ecapa is not in supports

yfchenlucky commented 6 months ago

We update the infer_sv.py and you can try again for ecapa-tdnn model. ResNet and Res2net will release next month.