CompiledProgram训练出错
Created by: aprilvkuo
python2.7 paddle 1.5.2 cpu 多核
A方式使用: with_data_parallel会报错。 B方式正常运行。 调用方式如下
exec_strategy = fluid.ExecutionStrategy()
build_strategy = fluid.compiler.BuildStrategy()
# A
compiled_prog = fluid.compiler.CompiledProgram(train_prog).with_data_parallel(loss_name=cost.name, build_strategy=build_strategy, exec_strategy=exec_strategy)
# B
compiled_prog = fluid.compiler.CompiledProgram(train_prog)
报错信息如下:
File "run_classifier.py", line 132, in train
return_numpy=True)
File "/home/guohongjie/tmp/paddle/paddle_release_home/python/lib/python2.7/site-packages/paddle/fluid/executor.py", line 664, in run
program._compile(scope, self.place)
File "/home/guohongjie/tmp/paddle/paddle_release_home/python/lib/python2.7/site-packages/paddle/fluid/compiler.py", line 376, in _compile
scope=self._scope)
File "/home/guohongjie/tmp/paddle/paddle_release_home/python/lib/python2.7/site-packages/paddle/fluid/compiler.py", line 347, in _compile_data_parallel
self._exec_strategy, self._build_strategy, self._graph)
paddle.fluid.core_avx.EnforceNotMet: at [/paddle/paddle/fluid/framework/ir/graph_helper.cc:364]
PaddlePaddle Call Stacks:
0 0x7f7eb24b3360p void paddle::platform::EnforceNotMet::Init<char const*>(char const*, char const*, int) + 352
1 0x7f7eb24b36d9p paddle::platform::EnforceNotMet::EnforceNotMet(std::__exception_ptr::exception_ptr, char const*, int) + 137
2 0x7f7eb3807e04p paddle::framework::ir::GraphNum(paddle::framework::ir::Graph const&) + 4132
3 0x7f7eb267389fp paddle::framework::ParallelExecutor::ParallelExecutor(std::vector<boost::variant<paddle::platform::CUDAPlace, paddle::platform::CPUPlace, paddle::platform::CUDAPinnedPlace, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_>, std::allocator<boost::variant<paddle::platform::CUDAPlace, paddle::platform::CPUPlace, paddle::platform::CUDAPinnedPlace, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_, boost::detail::variant::void_> > > const&, std::vector<std::string, std::allocator<std::string> > const&, std::string const&, paddle::framework::Scope*, std::vector<paddle::framework::Scope*, std::allocator<paddle::framework::Scope*> > const&, paddle::framework::details::ExecutionStrategy const&, paddle::framework::details::BuildStrategy const&, paddle::framework::ir::Graph*) + 5951
4 0x7f7eb253efb8p
5 0x7f7eb24e2f9ep
6 0x7f7f0b8c5df3p PyObject_Call + 67
7 0x7f7f0b8d49cdp
8 0x7f7f0b8c5df3p PyObject_Call + 67
9 0x7f7f0b933bafp
10 0x7f7f0b93046fp
11 0x7f7f0b8c5df3p PyObject_Call + 67
12 0x7f7f0b97b4a6p PyEval_EvalFrameEx + 15014
13 0x7f7f0b9810bdp PyEval_EvalCodeEx + 2061
14 0x7f7f0b97e345p PyEval_EvalFrameEx + 26949
15 0x7f7f0b97e460p PyEval_EvalFrameEx + 27232
16 0x7f7f0b9810bdp PyEval_EvalCodeEx + 2061
17 0x7f7f0b97e345p PyEval_EvalFrameEx + 26949
18 0x7f7f0b9810bdp PyEval_EvalCodeEx + 2061
19 0x7f7f0b97e345p PyEval_EvalFrameEx + 26949
20 0x7f7f0b9810bdp PyEval_EvalCodeEx + 2061
21 0x7f7f0b97e345p PyEval_EvalFrameEx + 26949
22 0x7f7f0b9810bdp PyEval_EvalCodeEx + 2061
23 0x7f7f0b9811f2p PyEval_EvalCode + 50
24 0x7f7f0b9a9f42p PyRun_FileExFlags + 146
25 0x7f7f0b9ab2d9p PyRun_SimpleFileExFlags + 217
26 0x7f7f0b9c100dp Py_Main + 3149
27 0x7f7f0abbebd5p __libc_start_main + 245
28 0x4007a1p