This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Converting mobileNetV3.ONNX model to TIDL .bin format



Facing the below issue while converting MobileNetV3.onnx file to TIDL .bin format.

ONNX opset version is 10

SDK - tidl_j7_01_03_00_11

Which onnx opset version is able to support Div and Reducemean operator.  Could you please help me to solve this issue.

TIDL logs:

:~/ti-processor-sdk-rtos-j721e-evm-07_01_00_11/tidl_j7_01_03_00_11/ti_dl/utils/tidlModelImport$ ./out/tidl_model_import.out /home/sithara/ti-processor-sdk-rtos-j721e-evm-07_01_00_11/tidl_j7_01_03_00_11/ti_dl/test/testvecs/config/import/msi/onnx/tidl_import_mobileNet_srikanth.txt 

ONNX Model (Proto) File  : ../../test/testvecs/models/msi/srikanth/mobilenetv3_W512_H320_C3_final.onnx  

TIDL Network File      : ../../test/testvecs/config/tidl_models/onnx/512x320_srikanth/tidl_net_mobilenet320x512.bin  

TIDL IO Info File      : ../../test/testvecs/config/tidl_models/onnx/512x320_srikanth/tidl_io_mobilenet320x512_  

Current ONNX OpSet Version   : 10  

 ONNX operator Div is not suported now..  By passing

 ONNX operator Div is not suported now..  By passing

 ONNX operator Div is not suported now..  By passing

Could not find const or initializer 775 !!!

Only float and INT64 tensor is suported 

 ONNX operator ReduceMean is not suported now..  By passing

Could not find const or initializer 726 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 678 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 760 !!!

Only float and INT64 tensor is suported 

 ONNX operator ReduceMean is not suported now..  By passing

 ONNX operator Div is not suported now..  By passing

 ONNX operator Div is not suported now..  By passing

 ONNX operator ReduceMean is not suported now..  By passing

Could not find const or initializer 711 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 663 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 629 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 581 !!!

Only float and INT64 tensor is suported 

 ONNX operator ReduceMean is not suported now..  By passing

 ONNX operator ReduceMean is not suported now..  By passing

Could not find const or initializer 614 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 566 !!!

Only float and INT64 tensor is suported 

 ONNX operator Div is not suported now..  By passing

 ONNX operator Div is not suported now..  By passing

 ONNX operator Div is not suported now..  By passing

Could not find const or initializer 490 !!!

Only float and INT64 tensor is suported 

 ONNX operator ReduceMean is not suported now..  By passing

Could not find const or initializer 444 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 399 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 475 !!!

Only float and INT64 tensor is suported 

 ONNX operator ReduceMean is not suported now..  By passing

 ONNX operator ReduceMean is not suported now..  By passing

Could not find const or initializer 429 !!!

Only float and INT64 tensor is suported 

Could not find const or initializer 384 !!!

Only float and INT64 tensor is suported 

Running tidl_optimizeNet 

tidl_model_import.out: tidl_import_common.cpp:164: void* my_malloc(int): Assertion `ptr != NULL' failed.