FunAudioLLM / CosyVoice

Multi-lingual large voice generation model, providing inference, training and deployment full-stack ability.
https://funaudiollm.github.io/
Apache License 2.0
5.84k stars 624 forks source link

各个模型下的campplus.onnx和resource.zip中的内容都是一样的且在运行demo和解压时会报错 #61

Closed ZeeChao closed 3 months ago

ZeeChao commented 3 months ago

我按照readme的方法安装搭建环境,下载模型。在运行

unzip resource.zip -d .

时,报错

Archive: resource.zip End-of-central-directory signature not found. Either this file is not a zipfile, or it constitutes one disk of a multi-part archive. In the latter case the central directory and zipfile comment will be found on the last disk(s) of this archive. unzip: cannot find zipfile directory in one of resource.zip or resource.zip.zip, and cannot find resource.zip.ZIP, period.

然后,我发现resource.zip中的内容是

version https://git-lfs.github.com/spec/v1 oid sha256:a6ac6a63997761ae2997373e2ee1c47040854b4b759ea41ec48e4e42df0f4d73 size 28303423

同时,我还发现,三个模型中的campplus.onnx文件中,写的也是

version https://git-lfs.github.com/spec/v1 oid sha256:a6ac6a63997761ae2997373e2ee1c47040854b4b759ea41ec48e4e42df0f4d73 size 28303423

因此,我在运行readme的demo时,报错

Traceback (most recent call last): File "run.py", line 5, in cosyvoice = CosyVoice('/workspace/CosyVoice/pretrained_models/CosyVoice-300M-SFT') File "/workspace/CosyVoice/cosyvoice/cli/cosyvoice.py", line 30, in init self.frontend = CosyVoiceFrontEnd(configs['get_tokenizer'], File "/workspace/CosyVoice/cosyvoice/cli/frontend.py", line 49, in init self.campplus_session = onnxruntime.InferenceSession(campplus_model, sess_options=option, providers=["CPUExecutionProvider"]) File "/opt/conda/envs/tts/lib/python3.8/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 419, in init self._create_inference_session(providers, provider_options, disabled_optimizers) File "/opt/conda/envs/tts/lib/python3.8/site-packages/onnxruntime/capi/onnxruntime_inference_collection.py", line 460, in _create_inference_session sess = C.InferenceSession(session_options, self._model_path, True, self._read_config_from_model) onnxruntime.capi.onnxruntime_pybind11_state.InvalidProtobuf: [ONNXRuntimeError] : 7 : INVALID_PROTOBUF : Load model from /workspace/CosyVoice/pretrained_models/CosyVoice-300M-SFT/campplus.onnx failed:Protobuf parsing failed.

ZeeChao commented 3 months ago

需要补充一下,我把demo中的第5行改成了

cosyvoice = CosyVoice('/workspace/CosyVoice/pretrained_models/CosyVoice-300M-SFT')

不知道上面的问题是否跟这个改动有关系。我之所以这样改是希望可以调用已经下载到本地的模型,而不是每次预测从线上下载模型。如果保持demo中原来的写法,似乎不会出现上面提到的错误

Helium-327 commented 2 months ago

我也有同样的问题,检查了git-lfs,和 protobuf的版本,都没有问题,但仍然会报错

zhaoshuyan commented 2 weeks ago

是DEMO上的模型的问题,通过魔搭直接下载模型来跑,是可以加载成功