Error when using with the HiFiGAN vocoder

[godspirit00]

I was trying to synthesize speech using the pretrained model with the HiFiGAN vocoder.

First I generated the mel file with
python gen_forward.py --input_text 'this is whatever you want it to be' melgan

Then, in the HiFiGAN vocoder repo, as stated in its README, I put the mel file in the test_mel_files, and run
python inference_e2e.py --checkpoint_file "./model/generator_v3"

And I got the following error:

Removing weight norm...
Traceback (most recent call last):
  File "inference_e2e.py", line 89, in <module>
    main()
  File "inference_e2e.py", line 85, in main
    inference(a)
  File "inference_e2e.py", line 49, in inference
    x = torch.FloatTensor(x).to(device)
ValueError: could not determine the shape of object type 'NpzFile'

It seems to have something to do with the mel file.

What am I missing? What should I do?
Thanks.

[cschaefer26]

Hi, true I just found that the current HifiGAN repo requests numpy files. Just updated the code. Try with :

python gen_forward.py --input_text 'this is whatever you want it to be' hifigan

[godspirit00]

Thank you for your reply!
So I downloaded the code and tried again, now it says:

Loading tts checkpoint model/forward_step80k.pt
Traceback (most recent call last):
  File "gen_forward.py", line 71, in <module>
    tts_model, config = load_forward_taco(checkpoint_path)
  File "gen_forward.py", line 21, in load_forward_taco
    tts_model = ForwardTacotron.from_config(config)
  File "D:\TTS\ForwardTacotron-master\models\forward_tacotron.py", line 276, in from_config
    return ForwardTacotron(**model_config)
TypeError: __init__() got an unexpected keyword argument 'pitch_emb_dims'

Using Wavernn vocoder would show the same error.

[cschaefer26]

Hi, it seems that you trained your model with an old version of the repo - I made some major changes that couldn't be backward compatible. You have two options: 1) retrain the model with the updated code (probably better quality), or 2) use the old code (you would need to revert it to the point before updating) and just replace the generate_forward.py with the newest version.

[godspirit00]

Solved!
Thanks a lot!