Description of reusable and customizable AI artifacts in the Collective Knowledge format (can be automatically installed, plugged to CK workflows and collaboratively optimized across diverse hardware/models/data sets/libraries):
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
.cm
ai-artifact
module
wfe
.ckr.json
.gitignore
CHANGES
CONTRIBUTIONS
COPYRIGHT
LICENSE
README.md

README.md

Collective Knowledge repository for AI

compatibility License

Introduction

This is an on-going community project to unify and automate AI research, share AI artifacts as customizable and reusable components with JSON API using open-source CK framework, continuously optimize and co-design the whole AI/SW/HW stack across diverse hardware/models/data sets/libraries, and help researchers adapt to a Cambrian AI/SW/HW explosion and technological chaos:

logo

You can browse reusable AI artifacts from this repository online.

Current AI artifacts are for Caffe, Caffe2, TensorFlow, MXNet, PyTorch and CNTK.

See cKnowledge.org/ai and cKnowledge.org/request for more details.

Coordination of development

Minimal installation

The minimal installation requires:

  • Python 2.7 or 3.3+ (limitation is mainly due to unitests)
  • Git command line client.

Linux/MacOS

You can install CK in your local user space simply as follows:

$ git clone http://github.com/ctuning/ck ck-master
$ export PATH=$PWD/ck-master/bin:$PATH
$ export PYTHONPATH=$PWD/ck-master:$PYTHONPATH

You can also install CK via pip with sudo to avoid setting up environment variables yourself (skip sudo on Windows);

$ sudo pip install ck

Windows

First you need to download and install a few dependencies from the following sites:

You can then install CK as follows:

 $ git clone https://github.com/ctuning/ck.git ck
 $ set PATH={CURRENT PATH}\ck\bin;%PATH%

Usage

Pull CK repo with all dependencies:

$ ck pull repo:ck-ai

$ ck ls ai-artifact

$ ck dashboard ai-artifact

Add new artifact description:

$ ck add ai-artifact:{AI alias}

Follow interactive instructions to add an artifact description.

Feel free to ask further questions via CK public mailing list.

Related Publications with long term vision

@inproceedings{cm:29db2248aba45e59:cd11e3a188574d80,
    url = {http://arxiv.org/abs/1506.06256},
    title = {{Collective Mind, Part II: Towards Performance- and Cost-Aware Software Engineering as a Natural Science.}},
    author = {Fursin, Grigori and Memon, Abdul and Guillon, Christophe and Lokhmotov, Anton},
    booktitle = {{18th International Workshop on Compilers for Parallel Computing (CPC'15)}},
    publisher = {ArXiv},
    year = {2015},
    month = January,
    pdf = {http://arxiv.org/pdf/1506.06256v1}
}

Next steps

We gradually add references about all existing CK artifacts including their installation, usage and optimization. The CK AI consortium is also working on a unification of all AI APIs and meta descriptions - join us!