-
I have a question, maybe we can replace the KPDetector with some pretrained face landmark detection networks directly for VoxCeleb Dataset。
Have you make similar experiments. Maybe you have some…
-
I used the kaldi/egs/dihard_2018/v2 recipe to make front-end processing,and get the MFCC feature for training X-vector model. Then I reproduce the code of training x-vector on pytorch, the x-vector mo…
-
How can I use the encoder manually, with the encoder_preprocess? Where I put my audio file ?
from encoder.preprocess import preprocess_librispeech, preprocess_voxceleb1, preprocess_voxceleb2
…
-
Hi !
I'm trying to run the speaker embedding pipeline on VoxCeleb2.
I'm using the up to date (git pulled on april 2) version of pyannote, I got the VoxCeleb corpus and the pyannote db voxceleb,…
-
Thanks for your great work.
I see the Voxceleb is used for training in your paper, but the weight is not shared for now.
So may I ask for uploading the trained weight file in Voxceleb, please.
Yo…
-
This your method to compute EER:
![image](https://user-images.githubusercontent.com/61141271/80098409-d36db080-853a-11ea-94ff-5cfee6e56764.png)
I have made a test and I got:
![image](https://user-i…
-
**Short description**
Checksums for oxford_iiit_pet dataset is changed.
**Environment information**
* Operating System: Google Colab
* Python version: 3.6.9 (default, Nov 7 2019, 10:44:02) \n[G…
-
Hello. Thank you for sharing the code!
I have my own test set for speaker verification (there are 2 different test sets, I tried both, the problem occurs in both). I am having some troubles with eva…
-
1. In paper you've mentioned,
> For each epoch, we sample a maximum of 100 utterances from each of the 5,994 identities
In code that propagates as value of nSpeakers args parameter for angula…
-
Hello, ty for code! Can model works without `LoadPldaParamsFromKaldi`? I train NeuralPLDA with embeddings from another model and it overfitted on 3-4 epoch. So, how many pairs should i have to train …