install.md

Installation

Get started

We test the installation with general requirements:

• Ubuntu 20.04 or newer
• Python >= 3.8
• CUDA >= 11.1
• PyTorch >= 1.9.1

There are two different ways to use oikit, i.e. stand-alone and import-as-package.

• The stand-alone will create an isolated conda env called: oakink, all the dependencies will be set up in this env.
  If you are new to OakInk, use stand-alone installation.
• The import-as-package will install oikit as a package in your current conda env.
  We suppose that python, cudatookit, and pytorch have already been installed.

stand-alone

$ conda env create -f environment.yaml
$ conda activate oakink
$ pip install -r requirements.txt

This environment provide you a base environment to load and visualize the OakInk dataset.

⚠️ In this case, you need to use the oikit inside the OakInk directory.

import-as-package (recommended)

In most cases, you want to use oikit in other conda env. To be able to import oikit, you need:

1. activate the destination env (we suppose that python, cudatookit, and pytorch have already been installed)
2. cd to the OakInk directory and run: pip install .

Get MANO asset

Get the MANO hand model mano_v1_2.zip from the MANO website.

1. click Download on the top menu, this requires register & login.
2. on the Download page, navigate to Models & Code section, and click Models & Code.

Unzip mano_v1_2.zip and copy it into the assets folder.