search

wwbin2017 / bailing

百聆 是一个类似GPT-4o的语音对话机器人，通过ASR+LLM+TTS实现，时延低至800ms，低配置也可运行，支持打断
MIT License
5 stars 2 forks source link

运行main.py遇到兼容性问题 #1

Open miaohuil opened 6 hours ago

miaohuil commented 6 hours ago

我现在用的是ffmpeg6.1.1。 我的config.yaml配置如下： name: 百聆（bailing） version: 1.0

logging: level: debug

interrupt: false

具体处理时选择的模块

selected_module: Recorder: RecorderPyAudio ASR: FunASR VAD: SileroVAD LLM: OpenAILLM TTS: MacTTS Player: CmdPlayer

Recorder: RecorderPyAudio: output_file: tmp/

ASR: FunASR:

model_dir: models/SenseVoiceSmall

model_dir: FunAudioLLM/SenseVoiceSmall
output_file: tmp/

VAD: SileroVAD: sampling_rate: 16000 threshold: 0.5 min_silence_duration_ms: 200

LLM: OpenAILLM: model_name: deepseek-chat url: https://api.deepseek.com api_key: sk-****

TTS: MacTTS: voice: Tingting output_file: tmp/ EdgeTTS: voice: zh-CN-XiaoxiaoNeural output_file: tmp/ GTTS: lang: zh output_file: tmp/ CosyvoiceTTS: output_file: tmp/ CHATTTS: output_file: tmp/

Player: CmdPlayer: null PyaudioPlayer: null

我已经用huggingface-cli命令登录了hf，运行main.py报错： Microsoft Windows [版本 10.0.19045.4894] (c) Microsoft Corporation。保留所有权利。

(bailing.env) C:\LLM-CLASS\开源项目测试\bailing>python main.py 2024-09-21 12:34:30,016 - torio._extension.utils - DEBUG - Loading FFmpeg6 2024-09-21 12:34:30,021 - torio._extension.utils - DEBUG - Failed to load FFmpeg6 extension. Traceback (most recent call last): File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 116, in _find_ffmpeg_extension ext = _find_versionsed_ffmpeg_extension(ffmpeg_ver) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 108, in _find_versionsed_ffmpeg_extension _load_lib(lib) File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 94, in _load_lib torch.ops.load_library(path) File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torch_ops.py", line 1295, in load_library
ctypes.CDLL(path) File "C:\Python312\Lib\ctypes__init.py", line 379, in init self._handle = _dlopen(self._name, mode) ^^^^^^^^^^^^^^^^^^^^^^^^^ FileNotFoundError: Could not find module 'C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio\lib\libtorio_ffmpeg6.pyd' (or one of its dependencies). Try using the full path with constructor syntax. 2024-09-21 12:34:30,027 - torio._extension.utils - DEBUG - Loading FFmpeg5 2024-09-21 12:34:30,078 - torio._extension.utils - DEBUG - Failed to load FFmpeg5 extension. Traceback (most recent call last): File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 116, in _find_ffmpeg_extension ext = _find_versionsed_ffmpeg_extension(ffmpeg_ver) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 108, in _find_versionsed_ffmpeg_extension _load_lib(lib) File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 94, in _load_lib torch.ops.load_library(path) File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torch_ops.py", line 1295, in load_library
ctypes.CDLL(path) File "C:\Python312\Lib\ctypes__init.py", line 379, in init self._handle = _dlopen(self._name, mode) ^^^^^^^^^^^^^^^^^^^^^^^^^ FileNotFoundError: Could not find module 'C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio\lib\libtorio_ffmpeg5.pyd' (or one of its dependencies). Try using the full path with constructor syntax. 2024-09-21 12:34:30,085 - torio._extension.utils - DEBUG - Loading FFmpeg4 2024-09-21 12:34:30,137 - torio._extension.utils - DEBUG - Failed to load FFmpeg4 extension. Traceback (most recent call last): File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 116, in _find_ffmpeg_extension ext = _find_versionsed_ffmpeg_extension(ffmpeg_ver) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 108, in _find_versionsed_ffmpeg_extension _load_lib(lib) File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 94, in _load_lib torch.ops.load_library(path) File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torch_ops.py", line 1295, in load_library
ctypes.CDLL(path) File "C:\Python312\Lib\ctypes\init.py", line 379, in init__ self._handle = _dlopen(self._name, mode) ^^^^^^^^^^^^^^^^^^^^^^^^^ FileNotFoundError: Could not find module 'C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio\lib\libtorio_ffmpeg4.pyd' (or one of its dependencies). Try using the full path with constructor syntax. 2024-09-21 12:34:30,144 - torio._extension.utils - DEBUG - Loading FFmpeg 2024-09-21 12:34:30,145 - torio._extension.utils - DEBUG - Failed to load FFmpeg extension. Traceback (most recent call last): File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 116, in _find_ffmpeg_extension ext = _find_versionsed_ffmpeg_extension(ffmpeg_ver) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torio_extension\utils.py", line 106, in _find_versionsed_ffmpeg_extension raise RuntimeError(f"FFmpeg{version} extension is not available.") RuntimeError: FFmpeg extension is not available. 2024-09-21 12:34:37,859 - root - INFO - new registry table has been added: preprocessor_classes 2024-09-21 12:34:39,060 - root - INFO - new registry table has been added: adaptor_classes 2024-09-21 12:34:39,699 - root - INFO - new registry table has been added: lid_predictor_classes {'model_dir': 'FunAudioLLM/SenseVoiceSmall', 'output_file': 'tmp/'} funasr version: 1.1.6. 2024-09-21 12:34:43,423 - root - INFO - download models from model hub: hf 2024-09-21 12:34:43,432 - urllib3.connectionpool - DEBUG - Starting new HTTPS connection (1): huggingface.co:443
2024-09-21 12:34:44,153 - urllib3.connectionpool - DEBUG - https://huggingface.co:443 "GET /api/models/FunAudioLLM/SenseVoiceSmall/revision/main HTTP/11" 200 1665 Fetching 29 files: 100%|█████████████████████████████████████████████████████████| 29/29 [00:00<00:00, 2901.80it/s] Detect model requirements, begin to install it: C:\Users\mhl.cache\huggingface\hub\models--FunAudioLLM--SenseVoiceSmall\snapshots\3eb3b4eeffc2f2dde6051b853983753db33e35c3\requirements.txt install model requirements successfully 2024-09-21 12:37:08,137 - root - INFO - Loading pretrained params from C:\Users\mhl.cache\huggingface\hub\models--FunAudioLLM--SenseVoiceSmall\snapshots\3eb3b4eeffc2f2dde6051b853983753db33e35c3\model.pt 2024-09-21 12:37:08,156 - root - INFO - ckpt: C:\Users\mhl.cache\huggingface\hub\models--FunAudioLLM--SenseVoiceSmall\snapshots\3eb3b4eeffc2f2dde6051b853983753db33e35c3\model.pt C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\funasr\train_utils\load_pretrained_model.py:38: FutureWarning: You are using torch.load with weights_only=False (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for weights_only will be flipped to True. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via torch.serialization.add_safe_globals. We recommend you start setting weights_only=True for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature. src_state = torch.load(path, map_location=map_location) 2024-09-21 12:37:09,109 - root - INFO - scope_map: ['module.', 'None'] 2024-09-21 12:37:09,110 - root - INFO - excludes: None 2024-09-21 12:37:09,243 - root - INFO - Loading ckpt: C:\Users\mhl.cache\huggingface\hub\models--FunAudioLLM--SenseVoiceSmall\snapshots\3eb3b4eeffc2f2dde6051b853983753db33e35c3\model.pt, status:
2024-09-21 12:37:09,263 - httpx - DEBUG - load_ssl_context verify=True cert=None trust_env=True http2=False 2024-09-21 12:37:09,265 - httpx - DEBUG - load_verify_locations cafile='C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\certifi\cacert.pem' SileroVAD {'sampling_rate': 16000, 'threshold': 0.5, 'min_silence_duration_ms': 200} Traceback (most recent call last): File "C:\LLM-CLASS\开源项目测试\bailing\main.py", line 59, in robot = robot.Robot(config_path) ^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing\robot.py", line 49, in init self.vad = vad.create_instance( ^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing\vad.py", line 70, in create_instance return cls(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing\vad.py", line 27, in init__ self.model = load_silero_vad() ^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\silero_vad\model.py", line 23, in load_silero_vad model = init_jit_model(model_file_path) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\silero_vad\utils_vad.py", line 171, in init_jit_model model = torch.jit.load(model_path, map_location=device) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\LLM-CLASS\开源项目测试\bailing\bailing.env\Lib\site-packages\torch\jit_serialization.py", line 163, in load cpp_module = torch._C.import_ir_module(cu, os.fspath(f), map_location, _extra_files, _restore_shapes) # type: ignore[call-arg] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ RuntimeError: open file failed because of errno 2 on fopen: No such file or directory, file path: C:\LLM-CLASS\开源 项目测试\bailing\bailing.env\Lib\site-packages\silero_vad\data\silero_vad.jit

miaohuil commented 5 hours ago

能提供您的config.yaml模版和ffmpeg的版本吗？