datduong

To start a docker with torch, I use

nvidia-docker run -e TERM=dumb -i -t kaixhin/torch 

Running a main script like main.lua gives error that cutorch is not found. Below is partial error message shown

Training a SPEN on the Data using tag_cmd.sh
tag_cmd.sh: 5: [: 0: unexpected operator
-gradient_clip 1.0 -optim_method adam -evaluation_frequency 25 -save_frequency 25 -adam_epsilon 1e-8 -batches_per_epoch 100 -learning_rate_decay 0.0 -learning_rate 0.001 -num_epochs 30 -training_mode pretrain_unaries

datduong

I use

$nvidia-docker run -e TERM=dumb -v /usr/local/cuda:/usr/local/cuda -i -t kaixhin/torch /bin/bash

and I was able to run luarocks install cutorch. But when testing the library, I get the error on cuda.

require 'cutorch'
THCudaCheck FAIL file=/tmp/luarocks_cutorch-scm-1-7401/cutorch/lib/THC/THCGeneral.c line=70 error=35 : CUDA driver version is insufficient for CUDA runtime version