OpenNMT Forum

CUDA runtime error

Hello

I got this error while training using GPU with the docker version of opennmt-lua.

THCudaCheck FAIL file=/root/torch-distro/extra/cutorch/lib/THC/THCGeneral.c line=70 error=35 : CUDA driver version is insufficient for CUDA runtime version

/root/torch/bin/luajit: ./onmt/utils/Cuda.lua:76: /root/torch/share/lua/5.1/trepl/init.lua:389: cuda runtime error (35) : CUDA driver version is insufficient for CUDA runtime version at /root/torch-distro/extra/cutorch/lib/THC/THCGeneral.c:70

stack traceback:

[C]: in function 'error'

./onmt/utils/Cuda.lua:76: in function 'init'

train.lua:300: in function 'main'

train.lua:338: in main chunk

[C]: in function 'dofile'

/root/torch/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk

[C]: at 0x00405d50

Can you help me with this? I’m just running the quick start command for training using the -gpuid option

Did you use nvidia-docker?

Yes I was using it but there was a missing option in the command. I had to include the --gpus option and that solved the problem.
Now the command is
docker run -it --gpus 3 opennmt/opennmt:latest
Thanks for the reply