search

astorfi / 3D-convolutional-speaker-recognition

:speaker: Deep Learning & 3D Convolutional Neural Networks for Speaker Verification
Apache License 2.0
781 stars 275 forks source link

About the data pipeline #6

Closed QimingZheng closed 6 years ago

QimingZheng commented 6 years ago

Hi astorfi: Can you show an example of how to prepare data for the enrollment stage? I met some problem in this stage, I process the data as the generation of data for development, but it doesn't work, the hdf5 format is somewhat annoying m can you show how to implement it just by using some random data as an example? Thanks a lot

astorfi commented 6 years ago

Enrollment stage is exactly similar to development stage similar to this code but with different speakers. So basically, using the background model created in the development stage, the new speakers' models will be generated in enrollment phase. So in enrollment stage, the data pipeline is exactly similar to development phase with different identities who will be used in enrollment and evaluation phases.

Thanks, Sina

JayyyWu commented 6 years ago

@astorfi Hi Sina, May I way the dimension of utterance_enrollment is (108,80,40,1) while utterance_train is (12,80,40,20)? What does the last dimension denote? Thanks!