使用trt推理时报错 - Error Code 3: API Usage Error (Parameter check failed - 非trt正常

[jaminegod]

问题确认 Search before asking

• 我已经查询历史issue，没有发现相似的bug。I have searched the issues and found no similar bug report.

Bug组件 Bug Component

No response

Bug描述 Describe the Bug

nvidia driver:

NVIDIA-Linux-x86_64-570.133.07.run

paddlepaddle docker image:

docker.1ms.run/paddlepaddle/paddle 3.0.0-gpu-cuda11.8-cudnn8.9-trt8.6 e96386e860a8 2 weeks ago 30.8GB

packages:

[λ codegee] /home pip list | grep paddle
paddledet 2.6.0
paddlepaddle-gpu 3.0.0
paddleslim 2.6.0

run code:

git clone https://github.com/PaddlePaddle/PaddleDetection.git
cd PaddleDetection/deploy/auto_compression/
python test_det.py --model_path ./models/rtdetr_hgnetv2_l_6x_coco/ --config ./configs/rtdetr_reader.yml --precision fp32 --use_trt True --use_dynamic_shape False

Error Info:
I0422 06:11:28.907433 169 fuse_pass_base.cc:55] --- detected 83 subgraphs
--- Running IR pass [trans_layernorm_fuse_pass]
--- Running IR pass [remove_padding_recover_padding_pass]
--- Running IR pass [delete_remove_padding_recover_padding_pass]
--- Running IR pass [dense_fc_to_sparse_pass]
--- Running IR pass [dense_multihead_matmul_to_sparse_pass]
--- Running IR pass [elementwise_groupnorm_act_pass]
--- Running IR pass [preln_elementwise_groupnorm_act_pass]
--- Running IR pass [groupnorm_act_pass]
--- Running IR pass [elementwiseadd_transpose_pass]
--- Running IR pass [tensorrt_subgraph_pass]
I0422 06:11:28.995630 169 tensorrt_subgraph_pass.cc:320] --- detect a sub-graph with 7 nodes
I0422 06:11:29.020102 169 tensorrt_subgraph_pass.cc:903] Prepare TRT engine (Optimize model structure, Select OP kernel etc). This process may cost a lot of time.
E0422 06:11:32.719125 169 helper.h:131] 3: [network.cpp::addConcatenation::856] Error Code 3: API Usage Error (Parameter check failed at: optimizer/api/network.cpp::addConcatenation::856, condition: (inputs[j]) != nullptr
)

C++ Traceback (most recent call last):
0 paddle_infer::Predictor::Predictor(paddle::AnalysisConfig const&)
1 std::unique_ptr<paddle::PaddlePredictor, std::default_deletepaddle::PaddlePredictor > paddle::CreatePaddlePredictor<paddle::AnalysisConfig, (paddle::PaddleEngineKind)2>(paddle::AnalysisConfig const&)
2 paddle::AnalysisPredictor::Init(std::shared_ptrpaddle::framework::Scope const&, std::shared_ptrpaddle::framework::ProgramDesc const&)
3 paddle::AnalysisPredictor::PrepareProgram(std::shared_ptrpaddle::framework::ProgramDesc const&)
4 paddle::AnalysisPredictor::OptimizeInferenceProgram()
5 paddle::inference::analysis::Analyzer::RunAnalysis(paddle::inference::analysis::Argument*)
6 paddle::inference::analysis::IrAnalysisPass::RunImpl(paddle::inference::analysis::Argument*)
7 paddle::inference::analysis::IRPassManager::Apply(std::unique_ptr<paddle::framework::ir::Graph, std::default_deletepaddle::framework::ir::Graph >)
8 paddle::framework::ir::Pass::Apply(paddle::framework::ir::Graph*) const
9 paddle::inference::analysis::TensorRtSubgraphPass::ApplyImpl(paddle::framework::ir::Graph*) const
10 paddle::inference::analysis::TensorRtSubgraphPass::CreateTensorRTOp(paddle::framework::ir::Node*, paddle::framework::ir::Graph*, std::vector<std::string, std::allocator<std::string > > const&, std::vector<std::string, std::allocator<std::string > >, bool) const
11 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*)
12 paddle::inference::tensorrt::OpConverter::ConvertBlock(paddle::framework::proto::BlockDesc const&, std::unordered_set<std::string, std::hash<std::string >, std::equal_to<std::string >, std::allocator<std::string > > const&, paddle::framework::Scope const&, paddle::inference::tensorrt::TensorRTEngine*)
13 paddle::inference::tensorrt::SplitOpConverter::operator()(paddle::framework::proto::OpDesc const&, paddle::framework::Scope const&, bool)

Error Message Summary:
FatalError: Segmentation fault is detected by the operating system.
[TimeInfo: *** Aborted at 1745302292 (unix time) try "date -d @1745302292" if you are using GNU date ***]
[SignalInfo: *** SIGSEGV (@0x8) received by PID 169 (TID 0x7f33eac5e740) from PID 8 ***]

Segmentation fault (core dumped)

复现环境 Environment

nvidia driver:

NVIDIA-Linux-x86_64-570.133.07.run

paddlepaddle docker image:

docker.1ms.run/paddlepaddle/paddle 3.0.0-gpu-cuda11.8-cudnn8.9-trt8.6 e96386e860a8 2 weeks ago 30.8GB

packages:

[λ codegee] /home pip list | grep paddle
paddledet 2.6.0
paddlepaddle-gpu 3.0.0
paddleslim 2.6.0

run code:

git clone https://github.com/PaddlePaddle/PaddleDetection.git
cd PaddleDetection/deploy/auto_compression/
python test_det.py --model_path ./models/rtdetr_hgnetv2_l_6x_coco/ --config ./configs/rtdetr_reader.yml --precision fp32 --use_trt True --use_dynamic_shape False

Bug描述确认 Bug description confirmation

• 我确认已经提供了Bug复现步骤、代码改动说明、以及环境信息，确认问题是可以复现的。I confirm that the bug replication steps, code change instructions, and environment information have been provided, and the problem can be reproduced.

是否愿意提交PR？ Are you willing to submit a PR?

• 我愿意提交PR！I'd like to help by submitting a PR!