Files mentioned in the tts.py from_local() are missing from the repo

#17
by segaa - opened

I see the following files mentioned in the tts.py from_local() function are missing from the repo:
t3_cfg.safetensors
tokenizer.json

So the from_local() function call fails with the following error:

Traceback (most recent call last):
  File "/test/chatterbox-turbo/example_tts.py", line 18, in <module>
    model = ChatterboxTTS.from_local(model_path, device)
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/test/chatterbox/src/chatterbox/tts.py", line 145, in from_local
    t3_state = load_file(ckpt_dir / "t3_cfg.safetensors")
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/test/miniconda3/envs/chatterbox/lib/python3.11/site-packages/safetensors/torch.py", line 313, in load_file
    with safe_open(filename, framework="pt", device=device) as f:
         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
FileNotFoundError: No such file or directory: "/test/models/ResembleAI-chatterbox-turbo/t3_cfg.safetensors"

Actually I see that I need to use ChatterboxTurboTTS from the The example audio is in stereo. instead, so please disregard my question.
It works for me when the code is structured as following:

from chatterbox.tts_turbo import ChatterboxTurboTTS
...
model_path = "/test/models/ResembleAI-chatterbox-turbo"
model = ChatterboxTurboTTS.from_local(model_path, device)
text="""
some text
"""
AUDIO_PROMPT_PATH = "Target_Voice.wav"
wav = model.generate(text, audio_prompt_path=AUDIO_PROMPT_PATH)
segaa changed discussion status to closed

Sign up or log in to comment