在jetson xavier nx上，采用Jetpack5.0.2：nv-jetson-cuda11.4-cudnn8.4.1-trt8.4.1-jetpack5.0.2-xavier推理库，trt_int8报错
zhizhongqu opened this issue · 1 comments
预测脚本：
python3 infer_resnet.py --model_file=./resnet50/inference.pdmodel --params_file=./resnet50/inference.pdiparams --run_mode=trt_int8
报错如下：
I0117 21:55:26.689865 23205 engine.cc:675] ====== engine info ======
terminate called after throwing an instance of 'phi::enforce::EnforceNotMet'
what(): (InvalidArgument) thread local var predictor_id_per_thread must be initialized to >= 0, but now predictor_id_per_thread = -1
[Hint: Expected predictor_id_per_thread > -1, but received predictor_id_per_thread:-1 <= -1:-1.] (at /home/paddle/data/xly/workspace/24116/Paddle/paddle/fluid/inference/tensorrt/engine.h:298)
C++ Traceback (most recent call last):
0 paddle::operators::TensorRTEngineOp::RunCalibration(paddle::framework::Scope const&, phi::Place const&) const::{lambda()#1}::operator()() const
1 paddle::operators::TensorRTEngineOp::PrepareTRTEngine(paddle::framework::Scope const&, paddle::inference::tensorrt::TensorRTEngine*) const
2 paddle::inference::tensorrt::OpConverter::ConvertBlockToTRTEngine(paddle::framework::BlockDesc*, paddle::framework::Scope const&, std::vector<std::string, std::allocator<std::string > > const&, std::unordered_set<std::string, std::hash<std::string >, std::equal_to<std::string >, std::allocator<std::string > > const&, std::vector<std::string, std::allocator<std::string > > const&, paddle::inference::tensorrt::TensorRTEngine*)
3 paddle::inference::tensorrt::TensorRTEngine::FreezeNetwork()
4 paddle::inference::tensorrt::TensorRTEngine::GetEngineInfo()
5 paddle::inference::tensorrt::TensorRTEngine::context()
Error Message Summary:
FatalError: Process abort signal is detected by the operating system.
[TimeInfo: *** Aborted at 1674021326 (unix time) try "date -d @1674021326" if you are using GNU date ***]
[SignalInfo: *** SIGABRT (@0x3e800005a85) received by PID 23173 (TID 0xffff31ccf1e0) from PID 23173 ***]
你好，麻烦更新到最新的develop之后看是否有相关问题。