dlib not using CUDA even after confirmation that it would during build

Question

I am attempting to get dlib to use CUDA on Ubuntu. I know that the hardware is capable. I already have CUDA and cuDNN.

I got dlib from the git repo and ran the following:

cd build
cmake .. -DCMAKE_PREFIX_PATH=/usr/local/cuda
cmake . --build

The build process said that dlib found cuda and cuDNN, and that dlib will use cuda, but when I run a python shell, and type

import dlib
dlib.DLIB_USE_CUDA

it returns False.

I am not sure why this happens. I am using python3 by the way. I also already installed dlib via pip3, and am not sure if I am using the wrong one.

Your question seems to be a duplicate of [this one](https://stackoverflow.com/questions/51943050/dlib-not-using-cuda) — Robert Crovella, Jul 04 '19 at 02:28

score 0 · Answer 1 · answered Jul 09 '19 at 06:42

0

From the dlib creator website:

answered Jul 09 '19 at 06:42

pedro_bb7

1 Answers1