object detection cpu启动失败
绿荫强鹰 发布于2018-11-05 11:15 浏览:54 回复:3
0
收藏

我是小白,在试着在无GPU服务器上搭建paaddlepaddle环境,按照https://github.com/PaddlePaddle/models/tree/develop/fluid/PaddleCV/object_detection上的例子中的,执行到python -u train.py --batch_size=64 --dataset='pascalvoc' --pretrained_model='pretrained/ssd_mobilenet_v1_coco/' --use_gpu False 命令时,程序出错,没有跑起来;希望大神们指示,错误日志如下:

----------- Configuration Arguments -----------
ap_version: 11point
batch_size: 64
data_dir: data/pascalvoc
dataset: pascalvoc
enable_ce: False
epoc_num: 120
image_shape: 3,300,300
learning_rate: 0.001
mean_BGR: 127.5,127.5,127.5
model_save_dir: model
parallel: True
pretrained_model: pretrained/ssd_mobilenet_v1_coco/
use_gpu: 0
------------------------------------------------
/usr/local/lib/python2.7/dist-packages/paddle/fluid/evaluator.py:71: Warning: The DetectionMAP is deprecated, because maintain a modified program inside evaluator cause bug easily, please use fluid.metrics.DetectionMAP instead.
% (self.__class__.__name__, self.__class__.__name__), Warning)
*** Aborted at 1541382160 (unix time) try "date -d @1541382160" if you are using GNU date ***
PC: @ 0x0 (unknown)
*** SIGILL (@0x7fd6a4935002) received by PID 24290 (TID 0x7fd6ff134700) from PID 18446744072175702018; stack trace: ***
@ 0x7fd6fed20390 (unknown)
@ 0x7fd6a4935002 paddle::framework::VisitDataType<>()
@ 0x7fd6a491d2ff paddle::operators::math::set_constant_with_place<>()
@ 0x7fd6a3ee899b paddle::operators::FillConstantOp::RunImpl()
@ 0x7fd6a48d0b58 paddle::framework::OperatorBase::Run()
@ 0x7fd6a3d1f0fb paddle::framework::Executor::RunPreparedContext()
@ 0x7fd6a3d1fee0 paddle::framework::Executor::Run()
@ 0x7fd6a3c408bd _ZZN8pybind1112cpp_function10initializeIZN6paddle6pybindL13pybind11_initEvEUlRNS2_9framework8ExecutorERKNS4_11ProgramDescEPNS4_5ScopeEibbE63_vIS6_S9_SB_ibbEINS_4nameENS_9is_methodENS_7siblingEEEEvOT_PFT0_DpT1_EDpRKT2_ENUlRNS_6detail13function_callEE1_4_FUNEST_
@ 0x7fd6a3c90cb4 pybind11::cpp_function::dispatcher()
@ 0x4cb2d4 PyEval_EvalFrameEx
@ 0x4c2e05 PyEval_EvalCodeEx
@ 0x4ca6b5 PyEval_EvalFrameEx
@ 0x4c2e05 PyEval_EvalCodeEx
@ 0x4caf42 PyEval_EvalFrameEx
@ 0x4c2e05 PyEval_EvalCodeEx
@ 0x4c2ba9 PyEval_EvalCode
@ 0x4f20ef (unknown)
@ 0x4eca72 PyRun_FileExFlags
@ 0x4eb1f1 PyRun_SimpleFileExFlags
@ 0x49e18a Py_Main
@ 0x7fd6fe965830 __libc_start_main
@ 0x49da19 _start
@ 0x0 (unknown)
非法指令 (核心已转储)

收藏
点赞
0
个赞
共3条回复 最后由TLTJ1223回复于2018-11-06 17:26
#4TLTJ1223回复于2018-11-06 17:26:15

另外还有一个解决办法:在本地机器源码编译并安装即可。

可以参考:https://github.com/PaddlePaddle/Paddle/issues/14275

0
#3189******80回复于2018-11-05 22:46:06
#2 徒尔为回复
[代码]

厉害厉害

0
#2徒尔为回复于2018-11-05 20:31:48
以下指令能检查Linux电脑是否支持AVX:

if cat /proc/cpuinfo | grep -i avx; then echo Yes; else echo No; fi

如果输出是No,就需要选择使用no-AVX的镜像
0
TOP
切换版块