Commits on Nov 9, 2024

1. Update Dockerfile with CUDA 11.8 …

   docker image里的cuda好像要升级一下，不然有这个报错: server log:2024-11-09 12:31:36 2024-11-09 20:31:36,085 INFO send inference response
   2024-11-09 12:31:36 
     0%|          | 0/1 [00:00<?, ?it/s]2024-11-09 20:31:36,093 INFO synthesis text 你好,我是通义生成式语音大模型,请问有什么可以帮您的吗?
   2024-11-09 12:31:44 
     0%|          | 0/1 [00:08<?, ?it/s]
   2024-11-09 12:31:44 2024-11-09 20:31:44,105 ERROR Exception iterating responses: cuFFT error: CUFFT_INTERNAL_ERROR
   2024-11-09 12:31:44 Traceback (most recent call last):
   2024-11-09 12:31:44   File "/opt/conda/lib/python3.10/site-packages/grpc/_server.py", line 589, in _take_response_from_response_iterator
   2024-11-09 12:31:44     return next(response_iterator), True
   2024-11-09 12:31:44   File "/opt/CosyVoice/CosyVoice/runtime/python/grpc/server.py", line 62, in Inference
   2024-11-09 12:31:44     for i in model_output:
   2024-11-09 12:31:44   File "/opt/CosyVoice/CosyVoice/runtime/python/grpc/../../../cosyvoice/cli/cosyvoice.py", line 61, in inference_sft
   2024-11-09 12:31:44     for model_output in self.model.tts(**model_input, stream=stream, speed=speed):
   2024-11-09 12:31:44   File "/opt/CosyVoice/CosyVoice/runtime/python/grpc/../../../cosyvoice/cli/model.py", line 191, in tts
   2024-11-09 12:31:44     this_tts_speech = self.token2wav(token=this_tts_speech_token,
   2024-11-09 12:31:44   File "/opt/CosyVoice/CosyVoice/runtime/python/grpc/../../../cosyvoice/cli/model.py", line 138, in token2wav
   2024-11-09 12:31:44     tts_speech, tts_source = self.hift.inference(speech_feat=tts_mel, cache_source=hift_cache_source)
   2024-11-09 12:31:44   File "/opt/conda/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
   2024-11-09 12:31:44     return func(*args, **kwargs)
   2024-11-09 12:31:44   File "/opt/CosyVoice/CosyVoice/runtime/python/grpc/../../../cosyvoice/hifigan/generator.py", line 410, in inference
   2024-11-09 12:31:44     generated_speech = self.decode(x=speech_feat, s=s)
   2024-11-09 12:31:44   File "/opt/CosyVoice/CosyVoice/runtime/python/grpc/../../../cosyvoice/hifigan/generator.py", line 350, in decode
   2024-11-09 12:31:44     s_stft_real, s_stft_imag = self._stft(s.squeeze(1))
   2024-11-09 12:31:44   File "/opt/CosyVoice/CosyVoice/runtime/python/grpc/../../../cosyvoice/hifigan/generator.py", line 334, in _stft
   2024-11-09 12:31:44     spec = torch.stft(
   2024-11-09 12:31:44   File "/opt/conda/lib/python3.10/site-packages/torch/functional.py", line 641, in stft
   2024-11-09 12:31:44     return _VF.stft(input, n_fft, hop_length, win_length, window,  # type: ignore[attr-defined]
   2024-11-09 12:31:44 RuntimeError: cuFFT error: CUFFT_INTERNAL_ERROR

   

   swang109 authored Nov 9, 2024
   Configuration menu

   • View commit details
   • Copy full SHA for d051212
   • Browse repository at this point

   Copy the full SHA

   d051212 View commit details

   Browse the repository at this point in the history