search

astorfi / 3D-convolutional-speaker-recognition

:speaker: Deep Learning & 3D Convolutional Neural Networks for Speaker Verification
Apache License 2.0
780 stars 274 forks source link

Enrollment and Evaluation Dataset Problem #10

Closed SameerGurjar closed 6 years ago

SameerGurjar commented 6 years ago

how the shape is obtained in utterance_enrollment in enrollment-evaluation_sample_dataset.hdf5 : no_of_sample x 80 x 40 x1

SameerGurjar commented 6 years ago

D:\sp_rec\feaext\0-input\create_hdf5\development_sample_dataset_speaker.hdf5 (File) '' Last modif.: 'Fri Oct 27 21:30:45 2017' Object Tree: / (RootGroup) '' /label_test (EArray(12,), shuffle, blosc(5)) '' /label_train (EArray(12,), shuffle, blosc(5)) '' /utterance_test (EArray(12, 80, 40, 20), shuffle, blosc(5)) '' /utterance_train (EArray(12, 80, 40, 20), shuffle, blosc(5)) ''

D:\sp_rec\feaext\0-input\create_hdf5\enrollment-evaluation_sample_dataset.hdf5 (File) '' Last modif.: 'Fri Mar 16 16:43:55 2018' Object Tree: / (RootGroup) '' /label_enrollment (EArray(108,), shuffle, blosc(5)) '' /label_evaluation (EArray(12,), shuffle, blosc(5)) '' /utterance_enrollment (EArray(108, 80, 40, 1), shuffle, blosc(5)) '' /utterance_evaluation (EArray(12, 80, 40, 1), shuffle, blosc(5)) ''

astorfi commented 6 years ago

@SameerGurjar Please read the associated paper: Text-Independent Speaker Verification Using 3D Convolutional Neural Networks