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.

TDA4VM: Meta aichitecuture models

Part Number: TDA4VM


Hi, TI

  As mentioned on TI official website, meta aichitecuture models like YOLOV3 has unsupported operators, However, it can be executed on the TDA4 platform. Why?

Thanks and Regards,

faw

  • Hi Tian,

     Will you please share the documentation that you are referring to. BTW, YOLOv3 is fully supported. There are sample configs for yolov3 in ti_dl\test\testvecs\config\import\public\onnx\tidl_import_yolo3.txt  that you can refer to. I am attaching the config file for reference.

    modelType          = 2
    numParamBits       = 8
    numFeatureBits     = 8
    quantizationStyle  = 3
    #quantizationStyle  = 2
    inputNetFile       = "../../test/testvecs/models/public/onnx/yolov3-10.onnx"
    outputNetFile      = "../../test/testvecs/config/tidl_models/onnx/tidl_net_yolo3.bin"
    outputParamsFile   = "../../test/testvecs/config/tidl_models/onnx/tidl_io_yolo3_"
    inDataNorm  = 1
    inMean = 0 0 0
    inScale = 0.003921568627 0.003921568627 0.003921568627
    inDataFormat = 1
    inWidth  = 416
    inHeight = 416 
    inNumChannels = 3
    numFrames = 1
    inData  =   "../../test/testvecs/config/detection_list.txt"
    perfSimConfig = ../../test/testvecs/config/import/device_config.cfg
    inElementType = 0
    #outDataNamesList = "convolution_output,convolution_output1,convolution_output2"
    metaArchType = 4
    metaLayersNamesList =  "../../test/testvecs/config/import/public/onnx/tidl_import_yolo3_metaarch.prototxt"
    postProcType = 2