运行ddk/samples/ai_benchmark/code里面resnet18出错

板卡使用 新手上路
,
1

详细描述的问题：

  1. 系统软件版本： x3_ubuntu_v1.1.5

  2. 问题涉及的技术领域： AI Benchmark

  3. 问题描述：运行ddk/samples/ai_benchmark/code里面resnet18出错

  4. 复现概率：必现

  5. 提供必要的问题日志：

  6. root@ubuntu:~/ptq/script/classification/resnet18# source ../../base_config.sh

  7. root@ubuntu:~/ptq/script/classification/resnet18# export SHOW_FPS_LOG=1

  8. root@ubuntu:~/ptq/script/classification/resnet18# ${app} --config_file=workflow_accuracy.json --log_level=2

  9. ../../aarch64/bin/example --config_file=workflow_accuracy.json --log_level=2

  10. I0000 00:00:00.000000 120171 vlog_is_on.cc:197] RAW: Set VLOG level for “*” to 2

  11. I0425 09:22:18.640105 120171 simple_example.cc:37] EXAMPLE_SYSTEM

  12. I0425 09:22:18.640313 120171 simple_example.cc:38] EXAMPLE_REPORT

  13. I0425 09:22:18.640341 120171 simple_example.cc:39] EXAMPLE_DETAIL

  14. [BPU_PLAT]BPU Platform Version(1.3.3)!

  15. [HBRT] set log level as 0. version = 3.15.17.0

  16. [DNN] Runtime version = 1.17.3_(3.15.17 HBRT)

  17. [A][DNN][packed_model.cpp:222](1682385738962) [HorizonRT] The model builder version = 1.15.5

  18. I0425 09:22:19.097177 120171 workflow_plugin.cc:30] WorkflowPlugin inited.

  19. I0425 09:22:19.097626 120171 input_plugin.cc:52] InputProducerPlugin start.

  20. I0425 09:22:19.097677 120171 workflow_plugin.cc:37] WorkflowPlugin start.

  21. terminate called after throwing an instance of 'I0425 09:22:19.098210 120171 output_plugin.cc:134] OutputConsumerPlugin start.

  22. std::invalid_argument’

  23. what(): stoi

  24. Aborted

  25. 软件上是否有做自定义修改：否

2

你好，该事例需要在pc端docker中编译

3

就是docker里面编译后，到device端运行报错，如运行这个ptq/script/classification/resnet18# ./accuracy.sh脚本啊！