Collective Knowledge platform Logo
1.3.1

Introduction

  • Introduction
    • Project overview
    • What is the Collective Knowledge project (CK)
    • Glossary
      • CK framework
      • CK portal
      • CK component
      • CK repository
      • Reproducible benchmarking pipeline and autotuning workflow
      • cBench
      • Live SOTA dashboards
      • Reproduced research papers
      • Portable CK solution
      • CK solutions for MLPerf
      • The meaning behind the “c”
    • Why CK?
    • Why open CK format?
    • Our goals and plans
  • Feedback and feature requests
  • Acknowledgments

Getting Started

  • cBench installation
    • Prerequisites
    • Linux
    • MacOS
    • Windows
    • Android (Linux host)
    • Docker
  • CK SDK installation
    • Prerequisites
    • Linux
    • MacOS
    • Windows
    • Android (Linux host)
    • Docker
  • MLPerf benchmark automation demo

User Guide

  • cBench commands
    • Setup and low-level access
      • Check the client version
      • Update client dependencies (CK components)
      • Setup the cKnowledge account
      • Test the login
      • Access the open cKnowledge API
      • Start the local server
    • CK components
      • Download components
      • Publish components
      • List versions of a given component
      • Open a cKnowledge web page with a given component
      • Delete components
    • cKnowledge dashboards
      • Initialize a cKnowledge graph
      • Push results
    • cKnowledge solutions
      • Download and initialize the existing solution
      • Run a solution
      • Benchmark a solution
      • Activate a virtual environment for the solution
      • List local solutions
      • Find local solutions
      • Delete local solutions
      • Create a new solution
  • CK commands
  • Practical use cases and tutorials
    • Share and reuse CK components similar to PyPI
    • Create customizable dashboards for live papers and collaborative experiments
    • Use cross-platform software detection plugins
    • Use cross-platform meta packages
    • Use portable workflows
    • Prepare portable cKnowledge solutions
  • How to contribute

Resources

  • CK Portal API
    • JSON Input
    • JSON Output
    • Available actions
      • Test login
      • List cKnowledge AI/ML/systems components
      • Download components
      • Get component info
      • Publish components
      • Get results
      • Publish results
  • cBench Python package
    • Submodules
    • cbench.client module
    • cbench.comm module
    • cbench.comm_min module
    • cbench.config module
    • cbench.graph module
    • cbench.main module
    • cbench.obj module
    • cbench.setup module
    • cbench.solution module
    • Module contents
  • CK Python package
Collective Knowledge platform
  • »
  • Search

© Copyright 2015-2020 Grigori Fursin and the cTuning foundation

Built with Sphinx using a theme provided by Read the Docs.