Qwen3-TTS GPU加速错误 #980
Description
按照文档：
默认配置为 CPU 模式以兼容所有电脑。如果您拥有 NVIDIA 显卡并安装了 CUDA，可以通过以下步骤开启加速（推理速度提升约10倍）：
右键点击对应的 .bat 文件，选择“编辑”。
删除文件末尾的 --device cpu --dtype float32 代码。
保存并重新运行。
但是会出现该错误：
Traceback (most recent call last):
File "<frozen runpy>", line 198, in _run_module_as_main
File "<frozen runpy>", line 88, in _run_code
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\qwen_tts\cli\demo.py", line 634, in <module>
raise SystemExit(main())
^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\qwen_tts\cli\demo.py", line 608, in main
tts = Qwen3TTSModel.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\qwen_tts\inference\qwen3_tts_model.py", line 112, in from_pretrained
model = AutoModel.from_pretrained(pretrained_model_name_or_path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\transformers\models\auto\auto_factory.py", line 604, in from_pretrained
return model_class.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\qwen_tts\core\models\modeling_qwen3_tts.py", line 1872, in from_pretrained
model = super().from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\transformers\modeling_utils.py", line 277, in _wrapper
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\transformers\modeling_utils.py", line 5048, in from_pretrained
) = cls._load_pretrained_model(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\transformers\modeling_utils.py", line 5432, in _load_pretrained_model
caching_allocator_warmup(model, expanded_device_map, hf_quantizer)
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\transformers\modeling_utils.py", line 6090, in caching_allocator_warmup
device_memory = torch_accelerator_module.mem_get_info(index)[0]
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\torch\cuda\memory.py", line 897, in mem_get_info
return torch.cuda.cudart().cudaMemGetInfo(device)
^^^^^^^^^^^^^^^^^^^
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\torch\cuda\__init__.py", line 501, in cudart
_lazy_init()
File "F:\index\qwen3tts-win-0124\runtime\Lib\site-packages\torch\cuda\__init__.py", line 417, in _lazy_init
raise AssertionError("Torch not compiled with CUDA enabled")
AssertionError: Torch not compiled with CUDA enabled
应该是qwen3tts-win-0124环境安装中没有GPU PyTorch的版本安装。