CosyVoice用5070TI训练输出语音报错

15failed to import ttsfrd, use WeTextProcessing instead

16D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\diffusers\models\lora.py:393: FutureWarning: `LoRACompatibleLinear` is deprecated and will be removed in version 1.0.0. Use of `LoRACompatibleLinear` is deprecated. Please switch to PEFT backend by installing PEFT: `pip install peft`.

17deprecate("LoRACompatibleLinear", "1.0.0", deprecation_message)

182025-04-24 10:20:34,437 INFO input frame rate=25

19D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\torch\nn\utils\weight_norm.py:28: UserWarning: torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.

20warnings.warn("torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.")

21Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.

22Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.

23D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\onnxruntime\capi\onnxruntime_inference_collection.py:69: UserWarning: Specified provider 'CUDAExecutionProvider' is not in available provider names.Available providers: 'AzureExecutionProvider, CPUExecutionProvider'

24warnings.warn(

252025-04-24 10:20:35,917 WETEXT INFO building fst for zh_normalizer ...

262025-04-24 10:20:35,917 INFO building fst for zh_normalizer ...

272025-04-24 10:20:52,653 WETEXT INFO done

282025-04-24 10:20:52,653 INFO done

292025-04-24 10:20:52,653 WETEXT INFO fst path: D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\tn\zh_tn_tagger.fst

302025-04-24 10:20:52,653 INFO fst path: D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\tn\zh_tn_tagger.fst

312025-04-24 10:20:52,653 WETEXT INFO D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\tn\zh_tn_verbalizer.fst

322025-04-24 10:20:52,653 INFO D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\tn\zh_tn_verbalizer.fst

332025-04-24 10:20:52,658 WETEXT INFO found existing fst: D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\tn\en_tn_tagger.fst

342025-04-24 10:20:52,658 INFO found existing fst: D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\tn\en_tn_tagger.fst

352025-04-24 10:20:52,658 WETEXT INFO D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\tn\en_tn_verbalizer.fst

362025-04-24 10:20:52,658 INFO D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\tn\en_tn_verbalizer.fst

372025-04-24 10:20:52,658 WETEXT INFO skip building fst for en_normalizer ...

382025-04-24 10:20:52,658 INFO skip building fst for en_normalizer ...

39D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\site-packages\torch\cuda\__init__.py:209: UserWarning:

40NVIDIA GeForce RTX 5070 Ti with CUDA capability sm_120 is not compatible with the current PyTorch installation.

41The current PyTorch install supports CUDA capabilities sm_50 sm_60 sm_61 sm_70 sm_75 sm_80 sm_86 sm_90.

42If you want to use the NVIDIA GeForce RTX 5070 Ti GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/

43

44warnings.warn(

45time="2025-04-24T10:20:56+08:00" level=error msg="Command execution failed:exit status 1"

46[INFO] 2025-04-24 10:20:54 - soundClone {'type': 'soundClone', 'seed': 0, 'speed': 1, 'text': '应群主邀请做个关于教育的分享', 'promptAudio': 'C:\\Users\\admin\\AppData\\Roaming\\aigcpanel\\data\\hub\\2025\\04\\24\\10\\18_21_35X97G43Xq.wav', 'promptText': '大家好，我是来自南京的选手，我叫史律。我今年是已经步入中年，但是我依然活力四射。', 'crossLingual': False}

47

480%| | 0/6 [00:00<?, ?it/s]

490%| | 0/6 [00:00<?, ?it/s]

50Traceback (most recent call last):

51File "D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\runpy.py", line 196, in _run_module_as_main

52return _run_code(code, main_globals, None,

53File "D:\model\aigcpanel\aigcpanelcosyvoice\_aienv\lib\runpy.py", line 86, in _run_code

54exec(code, run_globals)

55File "D:\model\aigcpanel\aigcpanelcosyvoice\aigcpanelrun.py", line 2, in <module>

56aigcpanelapp.run()

57File "aigcpanelapp.py", line 53, in aigcpanelapp.run

58File "D:\model\aigcpanel\aigcpanelcosyvoice\cosyvoice\cli\cosyvoice.py", line 85, in inference_zero_shot

59model_input = self.frontend.frontend_zero_shot(i, prompt_text, prompt_speech_16k, self.sample_rate)

60File "D:\model\aigcpanel\aigcpanelcosyvoice\cosyvoice\cli\frontend.py", line 166, in frontend_zero_shot

61speech_feat, speech_feat_len[:] = speech_feat[:, :2 * token_len], 2 * token_len

62RuntimeError: CUDA error: no kernel image is available for execution on the device

63CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.

64For debugging consider passing CUDA_LAUNCH_BLOCKING=1.

65Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.

大家有遇到过解决的吗？