search

ConsistencyVC / ConsistencyVC-voive-conversion

Using joint training speaker encoder with consistency loss to achieve cross-lingual voice conversion and expressive voice conversion
https://consistencyvc.github.io/ConsistencyVC-demo-page
MIT License
134 stars 22 forks source link

Uploaded model mistake #8

Closed gandolfxu closed 1 year ago

gandolfxu commented 1 year ago

The checkpoint sunder cvc-whispers-three-emo-loss are not right!

ConsistencyVC commented 1 year ago

Sorry, can you explain what happened, which one is not right?

gandolfxu commented 1 year ago

In the file whisperconvert_exp.py, the checkpoint name do not match that in the gdriver. Instead the checkpoint name of XVC and EVC are the same.

parser = argparse.ArgumentParser() parser.add_argument("--hpfile", type=str, default="./logs/cvc-whispers-three-emo-loss/config.json", help="path to json config file") parser.add_argument("--ptfile", type=str, default="./logs/cvc-whispers-three-emo-loss/G_cvc-whispers-three-emo-loss.pth", help="path to pth file") parser.add_argument("--outdir", type=str, default="output/60_exp_crosslingual_whispers-three-emo-loss", help="path to output dir")

Image_20230714162326

ConsistencyVC commented 1 year ago

ummm...I rename it wrong. But actually the checkpoint is right, just rename it is OK. I will revise it.

ConsistencyVC commented 1 year ago

Ok, I renamed it, if there is still any problem, just contact me

Fritskee commented 1 year ago

@ConsistencyVC When is the D_cvc model used actually? When I swap the models in whisperconvert_exp to the D_cvc model, it complains that the config is incorrect

ConsistencyVC commented 1 year ago

@ConsistencyVC When is the D_cvc model used actually? When I swap the models in whisperconvert_exp to the D_cvc model, it complains that the config is incorrect

That is the discriminator...