Hi,
When I try to test mobilenetV1_ssd from mlperf on TDA4VM（zenodo.org/.../ssd_mobilenet_v1_coco_2018_01_28.onnx）, which is claimed to reach 385 fps on TDA4VM in <Performance and efficiency benchmarking with TDA4 Edge AI processors>, I meet a problem in importing with tidl_tool as following, how to deal with unsupport layer when import model? Is there a way to extract subgraph from original model?
=======================================================================================
jacky.lin@CTOSvr:~/workspace/ti/edgeai-tidl-tools/tidl_tools$ ./tidl_model_import.out mobilenetV1_ssd/tidl_import_mobileNet1.txt
ONNX Model (Proto) File : ./mobilenetV1_ssd/mlperf_ssd_mobilenet_v1_coco_2018_01_28.onnx
TIDL Network File : ./mobilenetV1_ssd/tidl_net_ssd_mobilenet_1_224.bin
TIDL IO Info File : ./mobilenetV1_ssd/tidl_param_ssd_mobilenet_1_224.bin
Current ONNX OpSet Version : 11
ONNX operator Loop is not suported now.. By passing
ONNX operator Loop is not suported now.. By passing
ONNX operator Loop is not suported now.. By passing
ONNX operator Loop is not suported now.. By passing
...
Segmentation fault (core dumped)
=======================================================================================
The TIDL tool is build from latest:
https://github.com/TexasInstruments/edgeai-tidl-tools.git
The import config file (tidl_import_mobileNet1.txt)：
=======================================================================================
# Default - 0
randParams = 0
# 0: Caffe, 1: TensorFlow, 2: ONNX, Default - 0
modelType = 2
# 0: Fixed quantization By tarininng Framework, 1: Dyanamic quantization by TIDL, Default - 1
quantizationStyle = 1
# quantRoundAdd/100 will be added while rounding to integer, Default - 50
quantRoundAdd = 50
numParamBits = 8
inputNetFile = "./mobilenetV1_ssd/mlperf_ssd_mobilenet_v1_coco_2018_01_28.onnx"
inputParamsFile = "NA"
outputNetFile = "./mobilenetV1_ssd/tidl_net_ssd_mobilenet_1_224.bin"
outputParamsFile = "./mobilenetV1_ssd/tidl_param_ssd_mobilenet_1_224.bin"
preProcType = 2
sampleInData = "../test_data/airshow.jpg"
tidlStatsTool = "./PC_dsp_test_dl_algo.out"
inWidth = 224
inHeight = 224
inNumChannels = 3
=======================================================================================
