where to find the model(*.pth) to test the effect with my own input wav?

[yihe1003]

where to find the model(*.pth) to test the effect with my own input wav?

[yihe1003]

Thank you so much for your kindly help! There might be someting wrong during my model download. After re_download the model, the issue is solved. your model is very effective! Great work!

[haoheliu]

Thanks for your comments. It's a little bit hard to reproduce the error you encountered. It would be great if you can reproduce this error on the Colab for me to check. Also, I would suggest checking the following points:

1. Check if you have already downloaded the checkpoint file.
   Normally, the vocoder checkpoint should be present in
   ~/.cache/voicefixer/synthesis_module/44100/model.ckpt-1490000_trimed.pt
   And the voicefixer checkpoint should be present in
   ~/.cache/voicefixer/analysis_module/checkpoints/epoch=15_trimed_bn.ckpt
   The code will run the downloading process of these two checkpoints once you run the import:
   from voicefixer import VoiceFixer, Vocoder

2. Check if the downloaded file is correct. (Using md5 code to check)
   As shown in https://zenodo.org/record/5464142.
   The md5 of "epoch=15_trimed_bn.ckpt" is md5:0081967be3700f2eb132e0c0045fcf83 and "model.ckpt-1490000_trimed.pt" is md5:ac6de16561dd6042daf8bb3a806dc818

3. Make sure you have correctly pip installed voicefixer==0.0.12