search

mravanelli / SincNet

SincNet is a neural architecture for efficiently processing raw audio samples.
MIT License
1.13k stars 262 forks source link

EER is very high on dataset VoxCeleb1 #77

Open heyayun18188 opened 4 years ago

heyayun18188 commented 4 years ago

I test the performance of your pretrained model on the famous dataset VoxCeleb1(http://www.robots.ox.ac.uk/~vgg/data/voxceleb/). The EER I got is 30%. I used the compute_d_vector.py to get the d_vector of the audio and computed the cosine distance between 2 audios. Could you tell me why? Thanks a lot.

briverse17 commented 4 years ago

How did you compute the EER?

glwnogg commented 3 years ago

I test the performance of your pretrained model on the famous dataset VoxCeleb1(http://www.robots.ox.ac.uk/~vgg/data/voxceleb/). The EER I got is 30%. I used the compute_d_vector.py to get the d_vector of the audio and computed the cosine distance between 2 audios. Could you tell me why? Thanks a lot.

HI, I'm working on applying the sincnet on voxceleb and would you mind specifying how to handle the problem? Appreciate that!