Can't get attribute 'HParams' on <module 'utils'

[Kevinlinpr]

args: Namespace(speaker_model='shibing624/parrots-gpt-sovits-speaker-maimai', speaker_name='MaiMai', device='cpu', half=False, text='你好，欢迎来北京。welcome to the city.', lang='auto', output_path='output_audio.wav')
2024-03-17 01:11:46.818 | DEBUG | parrots.tts:init:302 - Use device: cpu
2024-03-17 01:11:49.862 | INFO | parrots.tts:init:316 - Load pretrained parrots speaker: shibing624/parrots-gpt-sovits-speaker-maimai
Fetching 6 files: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 6/6 [00:00<00:00, 62601.55it/s]
2024-03-17 01:11:50.537 | DEBUG | parrots.tts:init:329 - Reference speaker config: {'reference_wav': 'ref.wav', 'speaker': 'MaiMai', 'character': 'singing female anchor', 'reference_language': 'zh', 'reference_prompt': '那我们，唠也唠了这么久了唠了有十几分钟了我们要不来唱唱，唱唱歌，想听什么，今天想听什么。'}, loaded from /Users/kevinlinpr/.cache/huggingface/hub/models--shibing624--parrots-gpt-sovits-speaker-maimai/snapshots/369f6de40db8590be8eb1627d7f55fbbdb4fa63b/MaiMai/config.json
Traceback (most recent call last):
File "/Users/kevinlinpr/AI-Waifu-Vtuber/parrots_test.py", line 23, in
m = TextToSpeech(
^^^^^^^^^^^^^
File "/opt/homebrew/anaconda3/lib/python3.11/site-packages/parrots/tts.py", line 339, in init
sovits_dict = torch.load(sovits_model_path, map_location="cpu")
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/anaconda3/lib/python3.11/site-packages/torch/serialization.py", line 1026, in load
return _load(opened_zipfile,
^^^^^^^^^^^^^^^^^^^^^
File "/opt/homebrew/anaconda3/lib/python3.11/site-packages/torch/serialization.py", line 1438, in _load
result = unpickler.load()
^^^^^^^^^^^^^^^^
File "/opt/homebrew/anaconda3/lib/python3.11/site-packages/torch/serialization.py", line 1431, in find_class
return super().find_class(mod_name, name)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: Can't get attribute 'HParams' on <module 'utils' (<_frozen_importlib_external.NamespaceLoader object at 0x2d761c4d0>)>

[shibing624]

same to #27

[Kevinlinpr]

MaiMai的模型，经常出现口吃、语无伦次的情况，请问有什么优化的解决方案吗

[shibing624]

换其他模型，训练语料尽量跟预测语料内容分布一致，口语人声模型不适合读科技论文类，其他同理。

[Kevinlinpr]

我把参考语音换了一下，效果就好了很多

[shibing624]

可以，参考语言我是随机选了一个，推荐用其他更适合的。 :)