[GH-ISSUE #12] qwen_tts is not available. #10

New issue

Open

opened 2026-10-30 21:03:27 +00:00 by bsakaguchi · 2 comments

bsakaguchi commented

2026-10-30 21:03:27 +00:00

Owner

Originally created by @zwukong on GitHub (Jan 31, 2026).
Original GitHub issue: https://github.com/1038lab/ComfyUI-QwenTTS/issues/12

路径不对吧

File "M:\ComfyUI\312_cu128\ComfyUI\custom_nodes\ComfyUI-QwenTTS\AILab_QwenTTS.py", line 993, in generate
return _voice_clone_generate(
^^^^^^^^^^^^^^^^^^^^^^
File "M:\ComfyUI\312_cu128\ComfyUI\custom_nodes\ComfyUI-QwenTTS\AILab_QwenTTS.py", line 685, in _voice_clone_generate
model = _load_model("Base", model_size, device, precision, attention)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "M:\ComfyUI\312_cu128\ComfyUI\custom_nodes\ComfyUI-QwenTTS\AILab_QwenTTS.py", line 382, in _load_model
raise RuntimeError(
RuntimeError: qwen_tts is not available. Please install dependencies in your ComfyUI environment and check the package path. Import error: cannot import name 'Qwen3TTSTokenizerV1Config' from 'qwen_tts.core' (unknown location)

Originally created by @zwukong on GitHub (Jan 31, 2026). Original GitHub issue: https://github.com/1038lab/ComfyUI-QwenTTS/issues/12 路径不对吧 File "M:\ComfyUI\312_cu128\ComfyUI\custom_nodes\ComfyUI-QwenTTS\AILab_QwenTTS.py", line 993, in generate return _voice_clone_generate( ^^^^^^^^^^^^^^^^^^^^^^ File "M:\ComfyUI\312_cu128\ComfyUI\custom_nodes\ComfyUI-QwenTTS\AILab_QwenTTS.py", line 685, in _voice_clone_generate model = _load_model("Base", model_size, device, precision, attention) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "M:\ComfyUI\312_cu128\ComfyUI\custom_nodes\ComfyUI-QwenTTS\AILab_QwenTTS.py", line 382, in _load_model raise RuntimeError( **RuntimeError: qwen_tts is not available.** Please install dependencies in your ComfyUI environment and check the package path. Import error: cannot import name 'Qwen3TTSTokenizerV1Config' from 'qwen_tts.core' (unknown location)

bsakaguchi commented

2026-10-30 21:03:28 +00:00

Author

Owner

@viklen30 commented on GitHub (Feb 13, 2026):

RuntimeError: qwen_tts is not available. Please install dependencies in your ComfyUI environment and check the package path. Import error: cannot import name 'Qwen3TTSTokenizerV1Config' from 'qwen_tts.core' (unknown location)

@viklen30 commented on GitHub (Feb 13, 2026): RuntimeError: qwen_tts is not available. Please install dependencies in your ComfyUI environment and check the package path. Import error: cannot import name 'Qwen3TTSTokenizerV1Config' from 'qwen_tts.core' (unknown location)

bsakaguchi commented

2026-10-30 21:03:29 +00:00

Author

Owner

@Kozatiju commented on GitHub (Mar 14, 2026):

Hi everyone ! I had the same issue and Gemini really helped a lot with this. I'd copy paste his solution here :

After digging into the code and logging the hidden exceptions, I found out that the root cause is a silent crash happening during the core module load. It's caused by a syntax change in recent versions of transformers regarding the @check_model_inputs() decorator (TypeError: check_model_inputs() missing 1 required positional argument: 'func').

Here is the simple workaround to fix it and get the node working:

The Fix:
You need to comment out the check_model_inputs decorator and its import in the tokenizer modeling files.

Go to your ComfyUI custom nodes folder:
ComfyUI\custom_nodes\ComfyUI-QwenTTS\qwen_tts\core\tokenizer_12hz\

Open the file modeling_qwen3_tts_tokenizer_v2.py.

Around line 42, comment out the import:

Python

# from transformers.utils.generic import check_model_inputs

Search for @check_model_inputs() (usually right above the forward functions, around line 499) and comment them all out:

Python

# @check_model_inputs()

Save the files, completely restart ComfyUI, and run your prompt!

Hope this helps anyone stuck on this!

@Kozatiju commented on GitHub (Mar 14, 2026): Hi everyone ! I had the same issue and Gemini really helped a lot with this. I'd copy paste his solution here : After digging into the code and logging the hidden exceptions, I found out that the root cause is a silent crash happening during the core module load. It's caused by a syntax change in recent versions of transformers regarding the @check_model_inputs() decorator (TypeError: check_model_inputs() missing 1 required positional argument: 'func'). Here is the simple workaround to fix it and get the node working: The Fix: You need to comment out the check_model_inputs decorator and its import in the tokenizer modeling files. Go to your ComfyUI custom nodes folder: ComfyUI\custom_nodes\ComfyUI-QwenTTS\qwen_tts\core\tokenizer_12hz\ Open the file modeling_qwen3_tts_tokenizer_v2.py. Around line 42, comment out the import: Python `# from transformers.utils.generic import check_model_inputs` Search for @check_model_inputs() (usually right above the forward functions, around line 499) and comment them all out: Python `# @check_model_inputs()` Save the files, completely restart ComfyUI, and run your prompt! Hope this helps anyone stuck on this!

bsakaguchi referenced this issue

2026-10-30 21:03:38 +00:00

[PR #21] Add compatibility for Transformers 5.x #22