I currently incur in the following error when initializing the network with DriveWorks:
DefaultLogger: [23-02-2023 17:43:31] DefaultLogger: Driveworks exception thrown: DW_INVALID_HANDLE: Cannot cast to C++ handle, given instance is a nullptr, type=P15dwContextObject
terminate called after throwing an instance of ‘std::runtime_error’
what(): [2023-02-23 17:43:31] DW Error DW_INVALID_HANDLE executing DW function:
dwDNN_initializeTensorRTFromFile(&dnn, “/mnt/persistent/mfddFP16.bin”, nullptr, DW_PROCESSOR_TYPE_GPU, contextHandle)
The line of code which fails is:
CHECK_DW_ERROR(dwDNN_initializeTensorRTFromFile(&dnn, "/mnt/persistent/mfddFP16.bin", nullptr, DW_PROCESSOR_TYPE_GPU, contextHandle));
The engine was generated using
./usr/local/driveworks/tools/dnn/tensorRT_optimization starting from an onnx file.
Could you help in solving this issue?
Could you share DRIVE OS release, DW details?
1- Drive OS version 6.0.5
2- DW version 5.8
If there is something else you should know please let me know how to get the information.
could you help on this regard? Let me know in case you need anything else.
Could you confirm if you are using same object detector sample code and just replaced the model file path?
No, in the first place I just tried to load the engine, removing the rest of the code.
But now I tried with the original code by changing the model - generated with tensorRT_optimization - and I am getting a different error, as in the attachment.
error.txt (5.4 KB)
As per recent error log, the input video path is wrong. Please double check the input video path. Did you try with absolute path?(Autonomous Vehicle Development Platforms | NVIDIA Docs). Please share the complete used command and log.
in the first place I just tried to load the engine, removing the rest of the code
Does that include removal dwInitialize()? As the error indicates passed parameter is null.
I’m closing this topic due to there is no update from you for a period, assuming this issue was resolved.
If still need the support, please open a new topic. Thanks
Is this still an issue to support? Any result can be shared? Thanks