ENet

Install

Obtain ENet

git clone --recursive https://github.com/TimoSaemann/ENet.git
cd ENet/caffe-enet/docker

Config caffe-enet docker

vim Makefile:

cpu_standalone: standalone/cpu/Dockerfile
> $(DOCKER) build -t caffe-enet:cpu standalone/cpu

gpu_standalone: standalone/gpu/Dockerfile
 $(DOCKER) build -t caffe-enet:gpu standalone/gpu

FROM_GPU = "nvidia/cuda:8.0-cudnn5-devel-ubuntu16.04"
FROM_CPU = "ubuntu:16.04"

---

vim standalone/gpu/Dockerfile:

FROM nvidia/cuda:8.0-cudnn5-devel-ubuntu16.04

I think it is the same as

FROM nvidia/cuda:8.0-cudnn5-devel

---

vim standalone/cpu/Dockerfile:

FROM ubuntu:16.04

Build caffe-enet

make all
or
docker build -t caffe-enet:gpu standalone/gpu #or cpu

Check docker image:

~/Projects/caffe-enet-docker$ sudo docker image ls
REPOSITORY          TAG                            IMAGE ID            CREATED             SIZE
caffe-enet          gpu                            7e2d7b05a1ca        19 minutes ago      3.05GB
nvidia/cuda         8.0-cudnn5-devel-ubuntu16.04   38622aa43f13        5 days ago          1.85GB

Refer https://github.com/TimoSaemann/ENet/tree/master/Tutorial

Check version

ENet/caffe-enet/docker$ sudo docker run -ti caffe-enet:gpu caffe --v
ersion
caffe version 1.0.0

Run test:

 sudo nvidia-docker run -ti caffe-enet:gpu bash -c "cd /opt/caffe/build; make runtest"

Reference:

• CUDA docker
• ckhung/ENet
• TimoSaemann/ENet