Hello I got an error when I try to run this model.
Hello I got an error when I try to run this model with tools.webui module.
Below is the full error stack.
Traceback (most recent call last):
File "/usr/local/lib/python3.12/threading.py", line 1075, in _bootstrap_inner
self.run()
File "/usr/local/lib/python3.12/threading.py", line 1012, in run
self._target(*self._args, **self._kwargs)
File "/opt/fishaudio151/fish_speech/models/text2semantic/inference.py", line 911, in worker
model, decode_one_token = load_model(
^^^^^^^^^^^
File "/opt/fishaudio151/fish_speech/models/text2semantic/inference.py", line 676, in load_model
model: Union[NaiveTransformer, DualARTransformer] = BaseTransformer.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/fishaudio151/fish_speech/models/text2semantic/llama.py", line 376, in from_pretrained
config = BaseModelArgs.from_pretrained(str(path))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/opt/fishaudio151/fish_speech/models/text2semantic/llama.py", line 93, in from_pretrained
return cls(**data)
^^^^^^^^^^^
TypeError: DualARModelArgs.init() got an unexpected keyword argument 'attention_o_bias'
It seems like DualARModelArgs Class has unexpected keyword argument 'attention_o_bias'
Is there a different way to run this model compare to fish-speech-1.5 model?
I successfully deployed fish-speech-1.5 Model using tools.webui but I am having trouble with openaudio-s1-mini.
I tryied fish-speech version v1.5.1, v1.4.3, v1.4.0 (from Github https://github.com/fishaudio/fish-speech) to run this openaudio-s1-mini model but all failed with same error.
Please let me know how to run this model.
Fixed and running well.
attention_o_bias error fixed when I try the latest code (not the v1.5.1 tag version fish-speech, but the latest Git repo).
After that new errors came up which was no module found audiotools(https://github.com/descriptinc/audiotools) and dac(https://github.com/descriptinc/descript-audio-codec)
After I installed two more python packages from the above github, it worked fine.
For those who need.