gated-attention-reader
gated-attention-reader copied to clipboard
cuda out of memory
when I running on dataset dailymail/cnn, occur a problem saying that:" cuda runtime error (2) : out of memory at /py/conda-bld/pytorch_1493677666423/work/torch/lib/THC/generic/THCStorage.cu:66"
I use one gpu with 12G storage, I want to now what device you run on those two dataset? Or what can I do to optimize the operation?