Created by: mcl-stone

使用serving部署，使用debug脚本检测报错如下： [root@c99a526ec87f paddle]# python3.6 test_client2.py W0712 09:09:58.147349 1843 analysis_predictor.cc:133] Profiler is activated, which might affect the performance I0712 09:09:58.785828 1843 analysis_predictor.cc:872] MODEL VERSION: 1.7.1 I0712 09:09:58.785851 1843 analysis_predictor.cc:874] PREDICTOR VERSION: 1.8.2 I0712 09:09:58.786041 1843 analysis_predictor.cc:471] ir_optim is turned off, no IR pass will be executed --- Running analysis [ir_graph_build_pass] --- Running analysis [ir_graph_clean_pass] --- Running analysis [ir_analysis_pass] --- Running analysis [ir_params_sync_among_devices_pass] I0712 09:09:58.998006 1843 ir_params_sync_among_devices_pass.cc:41] Sync params from CPU to GPU --- Running analysis [adjust_cudnn_workspace_size_pass] --- Running analysis [inference_op_replace_pass] --- Running analysis [ir_graph_to_program_pass] I0712 09:09:59.093183 1843 analysis_predictor.cc:493] ======= optimize end ======= W0712 09:09:59.122561 1843 device_context.cc:252] Please NOTE: device: 0, CUDA Capability: 75, Driver API Version: 10.2, Runtime API Version: 10.0 W0712 09:09:59.126153 1843 device_context.cc:260] device: 0, cuDNN Version: 7.6. Traceback (most recent call last): File "test_client2.py", line 38, in fetch=["multiclass_nms"]) File "/usr/local/lib/python3.6/site-packages/paddle_serving_app/local_predict.py", line 127, in predict outputs = self.predictor.run(inputs) paddle.fluid.core_avx.EnforceNotMet:

---

C++ Call Stacks (More useful to developers):

0 std::string paddle::platform::GetTraceBackString<std::string const&>(std::string const&, char const*, int) 1 paddle::platform::EnforceNotMet::EnforceNotMet(std::string const&, char const*, int) 2 paddle::operators::ConvOp::GetExpectedKernelType(paddle::framework::ExecutionContext const&) const 3 paddle::framework::OperatorWithKernel::ChooseKernel(paddle::framework::RuntimeContext const&, paddle::framework::Scope const&, paddle::platform::Place const&) const 4 paddle::framework::OperatorWithKernel::RunImpl(paddle::framework::Scope const&, paddle::platform::Place const&, paddle::framework::RuntimeContext*) const 5 paddle::framework::OperatorWithKernel::RunImpl(paddle::framework::Scope const&, paddle::platform::Place const&) const 6 paddle::framework::OperatorBase::Run(paddle::framework::Scope const&, paddle::platform::Place const&) 7 paddle::framework::NaiveExecutor::Run() 8 paddle::AnalysisPredictor::Run(std::vector<paddle::PaddleTensor, std::allocatorpaddle::PaddleTensor > const&, std::vector<paddle::PaddleTensor, std::allocatorpaddle::PaddleTensor >*, int)

---

Python Call Stacks (More useful to users):

File "/usr/lib64/python2.7/site-packages/paddle/fluid/framework.py", line 2525, in append_op attrs=kwargs.get("attrs", None)) File "/usr/lib64/python2.7/site-packages/paddle/fluid/layer_helper.py", line 43, in append_op return self.main_program.current_block().append_op(*args, **kwargs) File "/usr/lib64/python2.7/site-packages/paddle/fluid/layers/nn.py", line 1405, in conv2d "data_format": data_format, File "/PaddleDetection/ppdet/modeling/backbones/resnet.py", line 162, in _conv_norm name=_name + '.conv2d.output.1') File "/PaddleDetection/ppdet/modeling/backbones/resnet.py", line 430, in c1_stage name=_name) File "/PaddleDetection/ppdet/modeling/backbones/resnet.py", line 451, in call res = self.c1_stage(res) File "/PaddleDetection/ppdet/modeling/architectures/faster_rcnn.py", line 89, in build body_feats = self.backbone(im) File "/PaddleDetection/ppdet/modeling/architectures/faster_rcnn.py", line 248, in test return self.build(feed_vars, 'test') File "tools/export_model.py", line 108, in main test_fetches = model.test(feed_vars) File "tools/export_model.py", line 125, in main()

---

Error Message Summary:

InvalidArgumentError: input and filter data type should be consistent [Hint: Expected input_data_type == filter_data_type, but received input_data_type:2 != filter_data_type:5.] at (/paddle/paddle/fluid/operators/conv_op.cc:173) [operator < conv2d > error] W0712 09:10:02.100517 1923 device_tracer.cc:53] Invalid timestamp occurred. Please try increasing the FLAGS_multiple_of_cupti_buffer_size.

-------------------------> Profiling Report <-------------------------

Place: All Time unit: ms Sorted by total time in descending order in the same thread

Total time: 303.444 Computation time Total: 201.3 Ratio: 66.3387% Framework overhead Total: 102.143 Ratio: 33.6613%

------------------------- GpuMemCpy Summary -------------------------

GpuMemcpy Calls: 172 Total: 83.4565 Ratio: 27.5031% GpuMemcpyAsync Calls: 3 Total: 3.2345 Ratio: 1.06593% GpuMemcpySync Calls: 169 Total: 80.222 Ratio: 26.4372%

------------------------- Event Summary -------------------------

Event Calls Total CPU Time (Ratio) GPU Time (Ratio) Min. Max. Ave. Ratio.
thread0::load 169 202.819 202.819352 (1.000000) 0.000000 (0.000000) 0.017028 164.441 1.20011 0.668392
thread0::GpuMemcpySync:CPU->GPU 169 80.222 40.694908 (0.507279) 39.527102 (0.492721) 0.011679 23.2058 0.474686 0.264372
thread0::conv2d 1 17.1275 17.127503 (1.000000) 0.000000 (0.000000) 17.1275 17.1275 17.1275 0.0564438
thread0::GpuMemcpyAsync:CPU->GPU 3 3.2345 1.820567 (0.562860) 1.413929 (0.437140) 0.011064 3.19608 1.07817 0.0106593
thread0::feed 3 0.040244 0.040244 (1.000000) 0.000000 (0.000000) 0.006155 0.025493 0.0134147 0.000132624