Has anyone actually tried pre-made model combo of VCTK VITS with VITS HiFiGan vocoder?

[skol101]

Here's the command
tts --text "This was the time when I was young and happy and what not used to go here and there, then every where I see my eyes never despise the horizon." --out_path vctk-vits-hfgan.wav --speaker_idx p267 --model_name tts_models/en/vctk/vits --vocoder_name vocoder_models/en/vctk/hifigan_v2

Here's result:

> tts_models/en/vctk/vits is already downloaded.
 > vocoder_models/en/vctk/hifigan_v2 is already downloaded.
 > Using model: vits
 > initialization of speaker-embedding layers.
 > Vocoder Model: hifigan
 > Generator Model: hifigan_generator
 > Discriminator Model: hifigan_discriminator
Removing weight norm...
 > Text: This was the time when I was young and happy and what not used to go here and there, then every where I see my eyes never despise the horizon.
 > Text splitted to sentences.
['This was the time when I was young and happy and what not used to go here and there, then every where I see my eyes never despise the horizon.']
Traceback (most recent call last):
  File "/home/sk/anaconda3/envs/tts_one/bin/tts", line 33, in <module>
    sys.exit(load_entry_point('TTS', 'console_scripts', 'tts')())
  File "/home/sk/work/TTS/TTS/bin/synthesize.py", line 257, in main
    wav = synthesizer.tts(args.text, args.speaker_idx, args.speaker_wav)
  File "/home/sk/work/TTS/TTS/utils/synthesizer.py", line 262, in tts
    waveform = self.vocoder_model.inference(vocoder_input.to(device_type))
  File "/home/sk/work/TTS/TTS/vocoder/models/gan.py", line 64, in inference
    return self.model_g.inference(x)
  File "/home/sk/anaconda3/envs/tts_one/lib/python3.8/site-packages/torch/autograd/grad_mode.py", line 28, in decorate_context
    return func(*args, **kwargs)
  File "/home/sk/work/TTS/TTS/vocoder/models/hifigan_generator.py", line 281, in inference
    return self.forward(c)
  File "/home/sk/work/TTS/TTS/vocoder/models/hifigan_generator.py", line 248, in forward
    o = self.conv_pre(x)
  File "/home/sk/anaconda3/envs/tts_one/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
    return forward_call(*input, **kwargs)
  File "/home/sk/anaconda3/envs/tts_one/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 301, in forward
    return self._conv_forward(input, self.weight, self.bias)
  File "/home/sk/anaconda3/envs/tts_one/lib/python3.8/site-packages/torch/nn/modules/conv.py", line 297, in _conv_forward
    return F.conv1d(input, weight, bias, self.stride,
RuntimeError: Given groups=1, weight of size [128, 80, 7], expected input[1, 126720, 11] to have 80 channels, but got 126720 channels instead

[erogol]

Vits has the vocoder embedded in the model. So you dont need to use a separate Vocoder model. It is also not compatible to be used like that. See the VITS paper for more info.