keras yolov3 onnx model can not parsers by tensorrt c++ api

[XuDongHecs]

[E] [TRT] Parameter check failed at: ../builder/Network.cpp::addInput::465, condition: isValidDims(dims) ERROR: ModelImporter.cpp:80 In function importInput: [8] Assertion failed: *tensor = importer_ctx->network()->addInput( input.name().c_str(), trt_dtype, trt_dims)

[jiafatom]

Can you try the model in model zoo here to see if it can work?

[araleqiu]

WARNING: ONNX model has a newer ir_version (0.0.5) than this parser was built against (0.0.3). [E] [TRT] Parameter check failed at: c:\p4sw\sw\gpgpu\MachineLearning\DIT\release\5.1\builder\Network.cpp::nvinfer1::Network::addInput::465, condition: isValidDims(dims) ERROR: ModelImporter.cpp:80 In function importInput: [8] Assertion failed: *tensor = importer_ctx->network()->addInput( input.name().c_str(), trt_dtype, trt_dims)

I face the same problem,i run by tensorrt c++ api,tensorrt version is TensorRT-5.1.5.0.Windows10.x86_64.cuda-9.0.cudnn7.5.I try the model in model Zoo,the same Error happen.

[XuDongHecs]

if you solve this problem. contact me！

---Original--- From: "araleqiu"<[email protected]> Date: Wed, Sep 11, 2019 11:46 AM To: "onnx/keras-onnx"<[email protected]>; Cc: "Author"<[email protected]>;"XuDongHecs"<[email protected]>; Subject: Re: [onnx/keras-onnx] keras yolov3 onnx model can not parsers by tensorrt c++ api (#231)

WARNING: ONNX model has a newer ir_version (0.0.5) than this parser was built against (0.0.3). [E] [TRT] Parameter check failed at: c:\p4sw\sw\gpgpu\MachineLearning\DIT\release\5.1\builder\Network.cpp::nvinfer1::Network::addInput::465, condition: isValidDims(dims) ERROR: ModelImporter.cpp:80 In function importInput: [8] Assertion failed: *tensor = importer_ctx->network()->addInput( input.name().c_str(), trt_dtype, trt_dims)

I face the same problem,i run by tensorrt c++ api,tensorrt version is TensorRT-5.1.5.0.Windows10.x86_64.cuda-9.0.cudnn7.5.I try the model in model Zoo,the same Error happen.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[jiafatom]

@XuDongHecs can you post this issue to this repo? seems more related there.