hwangCuc

icon company Communication University of China icon location beijing icon location speech synthesis CUC

Results 1 comments of hwangCuc

For model/prior.py _initial_sample, why the prob is calculated as from N(0,1)?

hi，I have met the same problem when I joined a vq encoder after posterior and prior encoder.The kl was 1e+4 and won't converged.Did you finish the job? > Thanks for...