generate_sequence_by_rnn_lm - Cuda Error: out of memory
Created by: lightsailpro
Trying to use the generate_sequence_by_rnn_lm model for sequence prediction. But I have large vocabulary size about half million (non word). The GPU card has 12GB ram. I got memory error even with small batch size 32. Can this model handle large dictionary size - max_word_num = 500000?
INFO 2017-08-15 12:57:05,733 train.py:95] dictionay size = 461546
I0815 12:57:05.739673 4302 Util.cpp:166] commandline: --use_gpu=True --trainer_count=1
I0815 12:57:07.165385 4302 GradientMachine.cpp:85] Initing parameters..
I0815 12:57:25.333416 4302 GradientMachine.cpp:92] Init parameters done.
[INFO 2017-08-15 12:57:25,333 train.py:78] start training...
[INFO 2017-08-15 12:57:28,632 train.py:58] Pass 0, Batch 0, Cost 257.150330, {'classification_error_evaluator': 1.0}
F0815 12:57:49.568514 4302 hl_cuda_device.cc:273] Check failed: cudaSuccess == cudaStat (0 vs. 2) Cuda Error: out of memory
*** Check failure stack trace: ***
@ 0x7f96cf30c8bd google::LogMessage::Fail()
@ 0x7f96cf310415 google::LogMessage::SendToLog()
@ 0x7f96cf30c3b3 google::LogMessage::Flush()
@ 0x7f96cf31192e google::LogMessageFatal::~LogMessageFatal()
@ 0x7f96cf2ce79c hl_malloc_device()
@ 0x7f96cf18fd06 paddle::GpuAllocator::alloc()
@ 0x7f96cf260666 paddle::PoolAllocator::alloc()
@ 0x7f96cf163614 paddle::GpuMemoryHandle::GpuMemoryHandle()
@ 0x7f96cf16d094 paddle::GpuMatrix::resize()
@ 0x7f96cf17db22 paddle::Matrix::resizeOrCreate()
@ 0x7f96cf09a1ad paddle::Layer::resetSpecifyOutput()
@ 0x7f96cf09a311 paddle::Layer::reserveOutput()
@ 0x7f96cf00e63f paddle::FullyConnectedLayer::forward()
@ 0x7f96cefdba10 paddle::NeuralNetwork::forward()
@ 0x7f96cefc9033 paddle::GradientMachine::forwardBackward()
@ 0x7f96cf2e7caa GradientMachine::forwardBackward()
@ 0x7f96cef609ee _wrap_GradientMachine_forwardBackward
@ 0x52714b PyEval_EvalFrameEx
@ 0x555551 PyEval_EvalCodeEx
@ 0x525560 PyEval_EvalFrameEx
@ 0x555551 PyEval_EvalCodeEx
@ 0x524338 PyEval_EvalFrameEx
@ 0x555551 PyEval_EvalCodeEx
@ 0x524338 PyEval_EvalFrameEx
@ 0x555551 PyEval_EvalCodeEx
@ 0x524338 PyEval_EvalFrameEx
@ 0x5247ea PyEval_EvalFrameEx
@ 0x567d14 (unknown)
@ 0x465bf4 PyRun_FileExFlags
@ 0x46612d PyRun_SimpleFileExFlags
@ 0x466d92 Py_Main
@ 0x7f96d0ceaf45 __libc_start_main
Aborted (core dumped)