fluid vgg16 flowers数据集显存占用过大（最新镜像）

[leanna62]

镜像版本：dzhwinter/benchmark:latest
运行命令：python vgg16.py --device GPU --data_set flowers --batch_size 64 （已拉取最新代码）
GPU信息：单卡显存：11439MiB
错误信息，看起来显存占用还是过大。
Traceback (most recent call last):
File "vgg16.py", line 182, in
main()
File "vgg16.py", line 156, in main
fetch_list=[avg_cost] + accuracy.metrics)
File "/usr/local/lib/python2.7/dist-packages/paddle/v2/fluid/executor.py", line 164, in run
self.executor.run(program.desc, scope, 0, True, True)
paddle.v2.fluid.core.EnforceNotMet: enforce allocating <= available failed, 10484135494 > 1370095360
at [/paddle/Paddle/paddle/platform/gpu_info.cc:89]
PaddlePaddle Call Stacks:
0 0x7f7a41602c28p paddle::platform::GpuMaxChunkSize() + 5080
1 0x7f7a40b7fdfcp void* paddle::memory::Allocpaddle::platform::CUDAPlace(paddle::platform::CUDAPlace, unsigned long) + 476
2 0x7f7a40b0240ep paddle::framework::Tensor::mutable_data(boost::variant<paddle::platform::CUDAPlace, paddle::platform::CPUPlace, 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_, boost::detail::variant::void_>, std::type_index) + 446
3 0x7f7a40b9bffap float* paddle::framework::Tensor::mutable_data(boost::variant<paddle::platform::CUDAPlace, paddle::platform::CPUPlace, 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_, boost::detail::variant::void_>) + 90
4 0x7f7a40e9e549p paddle::operators::GPUDropoutKernel<paddle::platform::CUDADeviceContext, float, float>::Compute(paddle::framework::ExecutionContext const&) const + 279
5 0x7f7a41573cf4p paddle::framework::OperatorWithKernel::Run(paddle::framework::Scope const&, boost::variant<paddle::platform::CUDAPlace, paddle::platform::CPUPlace, 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_, boost::detail::variant::void_> const&) const + 2084
6 0x7f7a40b86747p paddle::framework::Executor::Run(paddle::framework::ProgramDesc const&, paddle::framework::Scope*, int, bool, bool) + 983
7 0x7f7a40aec113p void pybind11::cpp_function::initialize<pybind11::cpp_function::initialize<void, paddle::framework::Executor, paddle::framework::ProgramDesc const&, paddle::framework::Scope*, int, bool, bool, pybind11::name, pybind11::is_method, pybind11::sibling>(void (paddle::framework::Executor::)(paddle::framework::ProgramDesc const&, paddle::framework::Scope, int, bool, bool), pybind11::name const&, pybind11::is_method const&, pybind11::sibling const&)::{lambda(paddle::framework::Executor*, paddle::framework::ProgramDesc const&, paddle::framework::Scope*, int, bool, bool)#1}, void, paddle::framework::Executor*, paddle::framework::ProgramDesc const&, paddle::framework::Scope*, int, bool, bool, pybind11::name, pybind11::is_method, pybind11::sibling>(pybind11::cpp_function::initialize<void, paddle::framework::Executor, paddle::framework::ProgramDesc const&, paddle::framework::Scope*, int, bool, bool, pybind11::name, pybind11::is_method, pybind11::sibling>(void (paddle::framework::Executor::)(paddle::framework::ProgramDesc const&, paddle::framework::Scope, int, bool, bool), pybind11::name const&, pybind11::is_method const&, pybind11::sibling const&)::{lambda(paddle::framework::Executor*, paddle::framework::ProgramDesc const&, paddle::framework::Scope*, int, bool, bool)#1}&&, void ()(paddle::framework::Executor, paddle::framework::ProgramDesc const&, paddle::framework::Scope*, int, bool, bool), pybind11::name const&, pybind11::is_method const&, pybind11::sibling const&)::{lambda(pybind11::detail::function_call&)#3}::_FUN(pybind11::detail::function_call) + 579
8 0x7f7a40ae9d64p pybind11::cpp_function::dispatcher(_object*, _object*, _object*) + 1236
9 0x4cad00p PyEval_EvalFrameEx + 28048
10 0x4c2705p PyEval_EvalCodeEx + 597
11 0x4ca088p PyEval_EvalFrameEx + 24856
12 0x4c2705p PyEval_EvalCodeEx + 597
13 0x4ca7dfp PyEval_EvalFrameEx + 26735
14 0x4c2705p PyEval_EvalCodeEx + 597
15 0x4c24a9p PyEval_EvalCode + 25
16 0x4f19efp
17 0x4ec372p PyRun_FileExFlags + 130
18 0x4eaaf1p PyRun_SimpleFileExFlags + 401
19 0x49e208p Py_Main + 1736
20 0x7f7a5ad33830p __libc_start_main + 240
21 0x49da59p _start + 41

[dzhwinter]

this problem is fixed in PaddlePaddle/Paddle#7443.
Please have a try after it merged.