Unable to convert function return value to a Python type!

[Rain5168]

Hi teams,

I have run your benchmark tool:  w_openvino_toolkit_mtl_23ww19 try to pre-test Resnet50 on MTL NPU, after step by step, The CMD show following erroe massage: 

How to solve it?

[Step 8/11] Querying optimal runtime parameters
[ INFO ] Model:
[ INFO ] CACHING_PROPERTIES: {'LOG_LEVEL': 'RO'}
[ INFO ] DEVICE_ID:
[ INFO ] EXECUTION_DEVICES: 3720
[ INFO ] INTERNAL_SUPPORTED_PROPERTIES: {'CACHING_PROPERTIES': 'RO'}
[ INFO ] MODEL_PRIORITY: Priority.MEDIUM
[ INFO ] NETWORK_NAME: Model2
[ INFO ] NPU_CREATE_EXECUTOR: 1
[ ERROR ] Unable to convert function return value to a Python type! The signature was
(self: openvino._pyopenvino.CompiledModel, property: str) -> object
TypeError: Failed to convert parameter to Python representation!

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "C:\Users\Public\w_openvino_toolkit_windows_2023.2.0.13089.cfd42bd2cb0_x86_64\python\openvino\tools\benchmark\main.py", line 451, in main
value = compiled_model.get_property(k)
TypeError: Unable to convert function return value to a Python type! The signature was
(self: openvino._pyopenvino.CompiledModel, property: str) -> object

[Hairul_Intel]

Hi Rain5168,

Thank you for reaching out to us.

Please provide the following information so that we can investigate this issue further:

• System info (processor, operating systems, etc.)
• Model files or its repository
• Commands used when encountering the error

Regards,

Hairul

[Rain5168]

Dear Hairul,

• System info (processor, operating systems, etc.)

          - CPU : Intel(R) Core(TM) Ultra 7 155H 3.80 GHz

          - RAM : 8GB

          - NPU driver : 31.0.100.1688

          - OS :  Windows 11 Pro 22H2

          - Python version :3.11.4

          - OV version : 2023.2

• Model files or its repository

         - I followed intel Document Number: 782543 step by steps and path setting as same as SOP in the document

• Commands used when encountering the error

         - when I was running NPU at throughput mode benchmark at command prompt(cmd) , the command is "benchmark_app -d NPU -m resnet-50-v1_5-sparse50.xml -t 180 -layout [NCHW] -ip f16 -op f16 -hint throughput -load_config npu_config.json" as intel's SOP

it always stopped at [Step 8/11], log information as following :

----------------------------------------------------------------------------------------------------------------------------------------------

C:\Users\Rain>C:\Users\Public\ov_23.2\Scripts\activate

(ov_23.2) C:\Users\Rain>C:\Users\Public\w_openvino_toolkit_windows_2023.2.0.13089.cfd42bd2cb0_x86_64\setupvars.bat
Python 3.11.4
[setupvars.bat] OpenVINO environment initialized

(ov_23.2) C:\Users\Rain>cd C:\Users\Public\

(ov_23.2) C:\Users\Public>copy C:\Users\Public\tools\ir_models\IRv11\20230309_vpu-models-mtl-por-ir_v11_ov_2022.3.0-9752fafe8eb\resnet-50-v1_5-sparse50\onnx\FP16-INT8\resnet-50-v1_5-sparse50.* .
C:\Users\Public\tools\ir_models\IRv11\20230309_vpu-models-mtl-por-ir_v11_ov_2022.3.0-9752fafe8eb\resnet-50-v1_5-sparse50\onnx\FP16-INT8\resnet-50-v1_5-sparse50.bin
Overwrite .\resnet-50-v1_5-sparse50.bin? (Yes/No/All): all
C:\Users\Public\tools\ir_models\IRv11\20230309_vpu-models-mtl-por-ir_v11_ov_2022.3.0-9752fafe8eb\resnet-50-v1_5-sparse50\onnx\FP16-INT8\resnet-50-v1_5-sparse50.xml
2 file(s) copied.

(ov_23.2) C:\Users\Public>benchmark_app -d NPU -m resnet-50-v1_5-sparse50.xml -t 180 -layout [NCHW] -ip f16 -op f16 -hint throughput -load_config npu_config.json
[Step 1/11] Parsing and validating input arguments
[ INFO ] Parsing input parameters
[Step 2/11] Loading OpenVINO Runtime
[ INFO ] OpenVINO:
[ INFO ] Build ................................. 2023.2.0-13089-cfd42bd2cb0-HEAD
[ INFO ]
[ INFO ] Device info:
[ INFO ] NPU
[ INFO ] Build ................................. 2023.2.0-2449-c8c8b709ea4-HEAD
[ INFO ]
[ INFO ]
[Step 3/11] Setting device configuration
[Step 4/11] Reading model files
[ INFO ] Loading model files
[ INFO ] Read model took 78.14 ms
[ INFO ] Original model I/O parameters:
[ INFO ] Model inputs:
[ INFO ] images (node: images) : f32 / [...] / [1,3,224,224]
[ INFO ] Model outputs:
[ INFO ] output (node: output) : f32 / [...] / [1,1000]
[Step 5/11] Resizing model to match image sizes and given batch
[ INFO ] Model batch size: 1
[Step 6/11] Configuring input of the model
[ INFO ] Model inputs:
[ INFO ] images (node: images) : f16 / [N,C,H,W] / [1,3,224,224]
[ INFO ] Model outputs:
[ INFO ] output (node: output) : f16 / [...] / [1,1000]
[Step 7/11] Loading the model to the device
[ INFO ] Compile model took 500.81 ms
[Step 8/11] Querying optimal runtime parameters
[ INFO ] Model:
[ INFO ] CACHING_PROPERTIES: {'LOG_LEVEL': 'RO'}
[ INFO ] DEVICE_ID:
[ INFO ] EXECUTION_DEVICES: 3720
[ INFO ] INTERNAL_SUPPORTED_PROPERTIES: {'CACHING_PROPERTIES': 'RO'}
[ INFO ] MODEL_PRIORITY: Priority.MEDIUM
[ INFO ] NETWORK_NAME: Model2
[ INFO ] NPU_CREATE_EXECUTOR: 1
[ ERROR ] Unable to convert function return value to a Python type! The signature was
(self: openvino._pyopenvino.CompiledModel, property: str) -> object
TypeError: Failed to convert parameter to Python representation!

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "C:\Users\Public\w_openvino_toolkit_windows_2023.2.0.13089.cfd42bd2cb0_x86_64\python\openvino\tools\benchmark\main.py", line 451, in main
value = compiled_model.get_property(k)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: Unable to convert function return value to a Python type! The signature was
(self: openvino._pyopenvino.CompiledModel, property: str) -> object

----------------------------------------------------------------------------------------------------------------------------------------------------

sincerely appreciate for your generous support. Thank you very much.

[Rain5168]

Hi teams,

Is there a further solution available for this issue?

or is there any additional information I need to provide?

Please let me know. Thank you. 

Regards,

Rain5168

[Hairul_Intel]

Hi Rain5168,

Apologies for the delay.

We just got feedback from the relevant teams regarding this issue. They confirmed that no issues was observed when using Benchmark_app C++ sample instead.

They will be fixing the Python sample and will notify you as soon as possible.

Regards,

Hairul

[Rain5168]

Hi Hairul,

As above,

Is there any fixed Python sample available for openvino benchmark testing?

Any good news, pls notify me ,thank you. 

Regards,

Rain5168,

[MILK_Yu]

Hi Rain5168
I suggest using the "OpenVINO Archives Includes NPU plugin (w_openvino_toolkit_windows_2023.2.0.13089.cfd42bd2cb0_x86_64.zip)" and runing python file that under the samples\python\benchmark\ .

Regards,

Milk

[Hairul_Intel]

Hi Rain5168,

Thank you for your patience.

To get the Python sample for benchmark testing, you will need to build the OpenVINO package from source (master branch) or wait for the next OpenVINO™ release.

Regards,

Hairul

[Hairul_Intel]

Hi Rain5168,

This thread will no longer be monitored since we have provided information. If you need any additional information from Intel, please submit a new question.

Regards,

Hairul