Hi,
I try to test edgeAI Apps on SK-TDA4VM.
So I using CPP based edgeAI Apps. It can run TFL/TVM(CL/OD/SS) models.
but CPP based edgeAI App can't run any ONNX model.
it shows below message:
root@j7-evm:/opt/edge_ai_apps/apps_cpp# ./bin/Release/app_edgeai ../configs/object_detection.yaml libtidl_onnxrt_EP loaded 0x1613a690 ERROR : artifacts_folder not a directoryParsing ONNX Model Segmentation fault (core dumped)
root@j7-evm:/opt/model_zoo# ls -l ONR-OD-8020-ssd-lite-mobv2-coco-512x512/ total 56 drwxr-xr-x 2 root root 4096 Mar 16 05:12 artifacts drwxr-xr-x 2 root root 4096 Mar 16 05:12 model -rw-r--r-- 1 30409 3355 4215 Mar 22 07:45 param.yaml -rw-r--r-- 1 30409 3355 39083 Dec 13 07:14 run.log root@j7-evm:/opt/model_zoo# ls -l ONR-OD-8020-ssd-lite-mobv2-coco-512x512/artifacts/ total 6232 -rw-r--r-- 1 30409 3355 1824 Dec 13 05:16 allowedNode.txt -rw-r--r-- 1 30409 3355 37256 Dec 13 06:23 boxeslabels_tidl_io_1.bin -rw-r--r-- 1 30409 3355 4573360 Dec 13 06:23 boxeslabels_tidl_net.bin -rw-r--r-- 1 30409 3355 1129 Dec 13 06:23 boxeslabels_tidl_net.bin.layer_info.txt -rw-r--r-- 1 30409 3355 1075945 Dec 13 06:23 boxeslabels_tidl_net.bin.svg -rw-r--r-- 1 30409 3355 25066 Dec 13 06:23 boxeslabels_tidl_net.bin_netLog.txt -rw-r--r-- 1 30409 3355 649453 Dec 13 05:16 runtimes_visualization.svg root@j7-evm:/opt/model_zoo#
Also i tested Python based EdgeAI Apps. it works well using any ONNX sample models.
I don't know that why cpp based app can't run ONNX model.
How can i fix this problem?
Regards,
Lee.
