Skip to content

Paddle模型尝试转换

Jump to bottom
SWHL edited this page Apr 6, 2022 · 2 revisions

✗(2022-04-06) openvino的mo转模型到IR格式

  • openvino-dev: 2022.1.0
  • 转换命令: 
    mo --input_model=resources/models/speedyspeech_csmsc/speedyspeech_csmsc.pdmodel
  • 报错信息: 
    E:\PythonProjects\RapidTTS2>mo --input_model=resources/models/speedyspeech_csmsc/speedyspeech_csmsc.pdmodel
Model Optimizer arguments:
Common parameters:
        - Path to the Input Model:      E:\PythonProjects\RapidTTS2\resources/models/speedyspeech_csmsc/speedyspeech_csmsc.pdmodel
        - Path for generated IR:        E:\PythonProjects\RapidTTS2\.
        - IR output name:       speedyspeech_csmsc
        - Log level:    ERROR
        - Batch:        Not specified, inherited from the model
        - Input layers:         Not specified, inherited from the model
        - Output layers:        Not specified, inherited from the model
        - Input shapes:         Not specified, inherited from the model
        - Source layout:        Not specified
        - Target layout:        Not specified
        - Layout:       Not specified
        - Mean values:  Not specified
        - Scale values:         Not specified
        - Scale factor:         Not specified
        - Precision of IR:      FP32
        - Enable fusing:        True
        - User transformations:         Not specified
        - Reverse input channels:       False
        - Enable IR generation for fixed input shape:   False
        - Use the transformations config file:  None
Advanced parameters:
        - Force the usage of legacy Frontend of Model Optimizer for model conversion into IR:   False       
        - Force the usage of new Frontend of Model Optimizer for model conversion into IR:      False       
OpenVINO runtime found in:      f:\miniconda3\envs\pytorch\lib\site-packages\openvino
OpenVINO runtime version:       2022.1.0-7019-cdb9bec7210-releases/2022/1
Model Optimizer version:        2022.1.0-7019-cdb9bec7210-releases/2022/1
[libprotobuf ERROR C:\j\workspace\private-ci\ie\build-windows-vs2019@3\b\repos\openvino\thirdparty\protobuf\protobuf\src\google\protobuf\message_lite.cc:133] Can't parse message of type "paddle.framework.proto.ProgramDesc" because it is missing required fields: (cannot determine missing fields for lite message)
[ ERROR ]  -------------------------------------------------
[ ERROR ]  ----------------- INTERNAL ERROR ----------------
[ ERROR ]  Unexpected exception happened.
[ ERROR ]  Please contact Model Optimizer developers and forward the following information:
[ ERROR ]  Check 'm_fw_ptr->ParseFromIstream(&pb_stream)' failed at C:\j\workspace\private-ci\ie\build-windows-vs2019@3\b\repos\openvino\src\frontends\paddle\src\input_model.cpp:315:
FrontEnd API failed with GeneralFailure: :
Model can't be parsed

[ ERROR ]  Traceback (most recent call last):
File "f:\miniconda3\envs\pytorch\lib\site-packages\openvino\tools\mo\main.py", line 533, in main
    ret_code = driver(argv)
File "f:\miniconda3\envs\pytorch\lib\site-packages\openvino\tools\mo\main.py", line 489, in driver        
    graph, ngraph_function = prepare_ir(argv)
File "f:\miniconda3\envs\pytorch\lib\site-packages\openvino\tools\mo\main.py", line 394, in prepare_ir    
    ngraph_function = moc_pipeline(argv, moc_front_end)
File "f:\miniconda3\envs\pytorch\lib\site-packages\openvino\tools\mo\moc_frontend\pipeline.py", line 29, in moc_pipeline
    input_model = moc_front_end.load(argv.input_model)
openvino.pyopenvino.GeneralFailure: Check 'm_fw_ptr->ParseFromIstream(&pb_stream)' failed at C:\j\workspace\private-ci\ie\build-windows-vs2019@3\b\repos\openvino\src\frontends\paddle\src\input_model.cpp:315:
FrontEnd API failed with GeneralFailure: :
Model can't be parsed


[ ERROR ]  ---------------- END OF BUG REPORT --------------
[ ERROR ]  -------------------------------------------------

✗(2022-04-06) openvino直接推理acoustic步骤下的pdmodel模型

  • openvino: 2022.1.0
  • OS: Windows 10 64位
  • 推理代码: 
    from openvino.runtime import Core

# openvino读取paddle模型
ie = Core()
paddle_model = ie.read_model(pdmodel_path)
compiled_model = ie.compile_model(model=paddle_model, device_name='CPU')
paddle_session = compiled_model.create_infer_request()
  • 报错信息: 
    E:\PythonProjects\RapidTTS2>python tts2.py
初始化前处理部分
frontend done!
初始化提取特征模型
[libprotobuf ERROR C:\j\workspace\private-ci\ie\build-windows-vs2019@3\b\repos\openvino\thirdparty\protobuf\protobuf\src\google\protobuf\message_lite.cc:133] Can't parse message of type "paddle.framework.proto.ProgramDesc" because it is missing required fields: (cannot determine missing fields for lite message)
Traceback (most recent call last):
File "tts2.py", line 36, in <module>
    am_predictor = SpeedySpeechAcoustic(pdmodel_path, pdiparam_path)
File "E:\PythonProjects\RapidTTS2\acoustic\speedyspeech_csmsc.py", line 28, in __init__
    paddle_model = ie.read_model(pdmodel_path)
RuntimeError: Check 'm_fw_ptr->ParseFromIstream(&pb_stream)' failed at C:\j\workspace\private-ci\ie\build-windows-vs2019@3\b\repos\openvino\src\frontends\paddle\src\input_model.cpp:315:
FrontEnd API failed with GeneralFailure: :
Model can't be parsed

✗(2022-04-06) Paddle2onnx转换acoustic模型到onnx

  • Paddle2onnx: 0.9.2
  • 转换代码: 
    paddle2onnx --model_dir resources/models/speedyspeech_csmsc \
            --model_filename resources/models/speedyspeech_csmsc/speedyspeech_csmsc.pdmodel \
            --params_filename resources/models/speedyspeech_csmsc/speedyspeech_csmsc.pdiparams \
            --save_file tmp \
            --opset_version 12
  • 报错信息: 
    NotImplementedError:
There's 2 ops are not supported yet
=========== conditional_block ===========
=========== while ===========
Clone this wiki locally