anhttran/3dmm_cnn

Script for fast installation on clean ubuntu 16.04

seva100 opened this issue ยท 7 comments

@anhttran, huge thanks for open-sourcing this.
I would like to share the installation script for all dependencies of this repository, which I prepared for convenient installation on a clean Ubuntu 16.04, Cuda 9.2 with Conda.
This is not an issue, rather an improvement which someone might find helpful, as I hope.

Installation instructions:

  1. Download the repository and recreate additional files, as described on the repository home page
  2. Download and install CUDA for Ubuntu 16.04 (9.2 is the current latest version):
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/cuda-repo-ubuntu1604_9.2.88-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1604_9.2.88-1_amd64.deb
sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/7fa2af80.pub
sudo apt-get update
sudo apt-get install cuda
  1. Download latest Anaconda or Miniconda
  2. Do
conda create -n=3dmm_cnn python=2.7.*
source activate 3dmm_cnn
  1. Download install.sh to the project folder, where all the files from repository are contained.
  2. Then
    bash install.sh

The script installs everything in a conda virtual environment 3dmm_cnn.

Amazing. It will help to solve a lot of installation issues.

I will check and share it to people. Many thanks for your sharing.

Many thanks, @seva100. We can create a wiki page on the github with his suggestion! Thanks!

I have updated the installation script: https://pastebin.com/xB1bMSVC
It now contains line 78, which I forgot to add initially. It is needed for Caffe to be seen after the installation.

Thanks! I replaced it on the wiki page!

@anhttran @iacopomasi, however, this script turns out to be not working for the Pose&Expr estimation script (demoCode/testModelPoseExpr.py), at least on my machine. Main issue seems to be from OpenCV versions conflict.
May I ask what is the OpenCV version required for the second script?

BTW, I would be very grateful to have a docker container provided by you. I can assist with that if needed.