problem with converting output

auspicious3000 / autovc

AutoVC: Zero-Shot Voice Style Transfer with Only Autoencoder Loss

https://arxiv.org/abs/1905.05879

MIT License

976 stars 207 forks source link

problem with converting output #72

Open todalex opened 3 years ago

todalex commented 3 years ago

hi i have trained a model and did all the required steps .at last i converted using my trained model and youre training metadata and the spectrograms of make_spect.py file and the output is some cut to pices voices mostly and i just have one good resualt that is from a speaker to himself can you help me with speaker embbeding and the .pkl file you create for conversion