tensorRT预测对elementwise_mul报错:
Created by: faninSM
报错error如下: 0 std::string paddle::platform::GetTraceBackStringstd::string(std::string&&, char const*, int) 1 paddle::inference::tensorrt::plugin::ElementWisePlugin::enqueue(int, void const* const*, void**, void*, CUstream_st*) 2 nvinfer1::rt::DefaultLayer::executeProfiled(nvinfer1::rt::CommonContext const&, nvinfer1::rt::ExecutionParameters const&) const 3 nvinfer1::rt::ExecutionContext::execute(int, void**) 4 nvinfer1::builder::calibrateEngine(nvinfer1::IInt8Calibrator&, nvinfer1::ICudaEngine&, std::unordered_map<std::string, float, std::hashstd::string, std::equal_tostd::string, std::allocator<std::pair<std::string const, float> > >&) 5 nvinfer1::builder::buildEngine(nvinfer1::CudaEngineBuildConfig&, nvinfer1::rt::HardwareContext const&, nvinfer1::Network const&) 6 nvinfer1::builder::Builder::buildCudaEngine(nvinfer1::INetworkDefinition&) 7 paddle::inference::tensorrt::TensorRTEngine::FreezeNetwork() 8 paddle::inference::tensorrt::OpConverter::ConvertBlockToTRTEngine(paddle::framework::BlockDesc*, paddle::framework::Scope const&, std::vector<std::string, std::allocatorstd::string > const&, std::unordered_set<std::string, std::hashstd::string, std::equal_tostd::string, std::allocatorstd::string > const&, std::vector<std::string, std::allocatorstd::string > const&, paddle::inference::tensorrt::TensorRTEngine*) 9 paddle::operators::TensorRTEngineOp::PrepareTRTEngine(paddle::framework::Scope const&, paddle::inference::tensorrt::TensorRTEngine*) const 10 paddle::operators::TensorRTEngineOp::RunCalibration(paddle::framework::Scope const&, paddle::platform::Place const&) const::{lambda()#1 (closed)}::operator()() const
Error Message Summary:
Error: Not implemented. at (/pr/Paddle/paddle/fluid/inference/tensorrt/plugin/elementwise_op_plugin.cu:72)
使用代码:fluid.layers.elementwise_mul(conv0, mask1,axis=2) 其中conv0,与mask1的shape分别为 : ('conv0', (-1L, 512L, 7L, 7L)) ('mask1', (7L, 7L))
paddle训练测试没有问题,但在tensorRT执行infer出现问题。 看了elementwise_op_plugin.cu的代码逻辑,也确实会报错。
版本信息: W0324 13:27:50.024972 2913 device_context.cc:237] Please NOTE: device: 0, CUDA Capability: 61, Driver API Version: 10.2, Runtime API Version: 9.0 W0324 13:27:50.027416 2913 device_context.cc:245] device: 0, cuDNN Version: 7.1. W0324 13:27:50.027452 2913 device_context.cc:271] WARNING: device: 0. The installed Paddle is compiled with CUDNN 7.5, but CUDNN version in your machine is 7.1, which may cause serious incompatible bug. Please recompile or reinstall Paddle with compatible CUDNN version.
gpu:p40
vertion.txt: GIT COMMIT ID: db40ee86 WITH_MKL: ON WITH_MKLDNN: OFF WITH_GPU: ON CUDA version: 9.0 CUDNN version: v7 WITH_TENSORRT: ON