[Bug] XTTS v2.0 finetuning - wrong checkpoint links

[rlenain]

Describe the bug

Hi there,

I believe that in the new XTTS v2.0 fine tuning recipe, there needs to be a change to the following lines:

TOKENIZER_FILE_LINK = "https://coqui.gateway.scarf.sh/hf-coqui/XTTS-v1/v2.0/vocab.json"
XTTS_CHECKPOINT_LINK = "https://coqui.gateway.scarf.sh/hf-coqui/XTTS-v1/v2.0/model.pth"

It's impossible to reach these URLs.

Thanks.

To Reproduce

python recipes/ljspeech/xtts_v2/train_gpt_xtts.py

Expected behavior

Training

Logs

/home/raph/repos/TTS/TTS/tts/layers/xtts/trainer/dataset.py:10: UserWarning: Torchaudio's I/O functions now support par-call bakcend dispatch. Importing backend implementation directly is no longer guaranteed to work. Please use `backend` keyword with load/save/info function, instead of calling the udnerlying implementation directly.
  from torchaudio.backend.soundfile_backend import load as torchaudio_soundfile_load
/home/raph/repos/TTS/TTS/tts/layers/xtts/trainer/dataset.py:11: UserWarning: Torchaudio's I/O functions now support par-call bakcend dispatch. Importing backend implementation directly is no longer guaranteed to work. Please use `backend` keyword with load/save/info function, instead of calling the udnerlying implementation directly.
  from torchaudio.backend.sox_io_backend import load as torchaudio_sox_load
/home/raph/miniconda3/envs/TTS/lib/python3.10/site-packages/torch/nn/utils/weight_norm.py:30: UserWarning: torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.
  warnings.warn("torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.")
Traceback (most recent call last):
  File "/home/raph/repos/TTS/recipes/ljspeech/xtts_v2/train_gpt_xtts.py", line 232, in <module>
    main()
  File "/home/raph/repos/TTS/recipes/ljspeech/xtts_v2/train_gpt_xtts.py", line 204, in main
    model = GPTTrainer.init_from_config(config)
  File "/home/raph/repos/TTS/TTS/tts/layers/xtts/trainer/gpt_trainer.py", line 500, in init_from_config
    return GPTTrainer(config)
  File "/home/raph/repos/TTS/TTS/tts/layers/xtts/trainer/gpt_trainer.py", line 79, in __init__
    self.xtts.tokenizer = VoiceBpeTokenizer(self.args.tokenizer_file)
  File "/home/raph/repos/TTS/TTS/tts/layers/xtts/tokenizer.py", line 540, in __init__
    self.tokenizer = Tokenizer.from_file(vocab_file)
Exception: expected value at line 1 column 1
 ~/repos/TTS  main !1 ?3  vim recipes/ljspeech

### Environment

```shell
{
    "CUDA": {
        "GPU": [
            "NVIDIA A100-PCIE-40GB",
            "NVIDIA A100-PCIE-40GB",
            "NVIDIA A100-PCIE-40GB",
            "NVIDIA A100-PCIE-40GB"
        ],
        "available": true,
        "version": "12.1"
    },
    "Packages": {
        "PyTorch_debug": false,
        "PyTorch_version": "2.1.0+cu121",
        "TTS": "0.20.0",
        "numpy": "1.22.0"
    },
    "System": {
        "OS": "Linux",
        "architecture": [
            "64bit",
            "ELF"
        ],
        "processor": "x86_64",
        "python": "3.10.13",
        "version": "#98-Ubuntu SMP Mon Oct 2 15:18:56 UTC 2023"
    }
}

Additional context

No response

[AWAS666]

The most up to date links seem to be in models.json.

"https://coqui.gateway.scarf.sh/hf-coqui/XTTS-v2/main/model.pth",
"https://coqui.gateway.scarf.sh/hf-coqui/XTTS-v2/main/config.json",
"https://coqui.gateway.scarf.sh/hf-coqui/XTTS-v2/main/vocab.json",
"https://coqui.gateway.scarf.sh/hf-coqui/XTTS-v2/main/hash.md5"

as a quick fix, I'll create a PR for it.

[Edresson]

It was fixed on #3149 . However, currently, XTTS v2.0 fine-tuning is not supported. It uses a new DVAE that is not implemented. We are working to fix this issue soon as possible.

[Edresson]

The PR #3154 fixed this issue.

[Yaodada12]

It was fixed on #3149 . However, currently, XTTS v2.0 fine-tuning is not supported. It uses a new DVAE that is not implemented. We are working to fix this issue soon as possible.

Can xtts v2.0 fine-tuning on a character's audio like RVC to achieve better performance?