Can't create engine from model files yolov3

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

Hey guys, I got an error when try to ‘deepstream-app -c deepstream_app_config_yoloV3.txt’ and perform a deepsream-python-app:

desktop:/opt/nvidia/deepstream/deepstream-5.0/sources/objectDetector_Yolo$ deepstream-app -c deepstream_app_config_yoloV3.txt
Unknown or legacy key specified ‘is-classifier’ for group [property]
Warn: ‘threshold’ parameter has been deprecated. Use ‘pre-cluster-threshold’ instead.

Using winsys: x11
gstnvtracker: Loading low-level lib at /opt/nvidia/deepstream/deepstream-5.0/lib/libnvds_mot_klt.so
gstnvtracker: Optional NvMOT_RemoveStreams not implemented
gstnvtracker: Batch processing is OFF
gstnvtracker: Past frame output is OFF
0:00:00.246684862 9116 0x30718690 INFO nvinfer gstnvinfer.cpp:619:gst_nvinfer_logger:<primary_gie> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::buildModel() <nvdsinfer_context_impl.cpp:1715> [UID = 1]: Trying to create engine from model files
Loading pre-trained weights…
Loading weights of yolov3 complete!
Total Number of weights read : 62001757
terminate called after throwing an instance of ‘std::out_of_range’
what(): map::at
Aborted (core dumped)

key point:
Trying to create engine from model files
Loading pre-trained weights…
Loading weights of yolov3 complete!
Total Number of weights read : 62001757
terminate called after throwing an instance of ‘std::out_of_range’
** what(): map::at**
Aborted (core dumped)

I did the same work before and they were fine,
now I use a new device with JP4.4.1, deepstream5.0.1, tensorrt7.1.3,
how to fix the problem?

thx
cy

5

which deepstream version and device you used, when this work?
which new device you are using for deesptream 5.0.1, this did not work?
I tried with original yolov3 model, it works with deepstream 5.0.1
just paste some log,

0:00:00.216752059 27961 0x13fc5930 INFO nvinfer gstnvinfer.cpp:619:gst_nvinfer_logger:<primary_gie> NvDsInferContext[UID 1]: Info from NvDsInferContextImpl::buildModel() <nvdsinfer_context_impl.cpp:1715> [UID = 1]: Trying to create engine from model files
Loading pre-trained weights…
Loading weights of yolov3 complete!
Total Number of weights read : 62001757
Loading pre-trained weights…
Loading weights of yolov3 complete!
Total Number of weights read : 62001757
Building Yolo network…
layer inp_size out_size weightPtr
(0) conv-bn-leaky 3 x 608 x 608 32 x 608 x 608 992
(1) conv-bn-leaky 32 x 608 x 608 64 x 304 x 304 19680
your error should be from sources/objectDetector_Yolo/nvdsinfer_custom_impl_Yolo/trt_utils.cpp function loadWeights returned
one more, did you get this issue with your customized yolo model or original yolo model?

6

I got a new jetson xavier NX, I used another xavier NX with deepstream5.0 and jp4.4DP before.

I got yolov3.weights and cfg by ‘prebuild.sh’. I tried resnet10.caffemodel in samples and it’s fine.

so, weights problem is more likely?

7

problem solved.
stupid mistake.
yolov3.cfg mismatch.
thank you.