Skip to main content
Version: 0.8.x [Latest Beta]

Integration overview

This chapter explains how to install, configure, and operate the DENKflow SDK in real user environments from Python and C/C++.

Who this chapter is for​

Use this chapter if you want to:

  • run exported .denkflow pipelines from the DENKweit Vision AI Hub
  • pick the right runtime for your hardware
  • deploy on Intel, NVIDIA, or Windows GPU systems
  • package your application in Docker
  • build custom pipelines from .denkmodel files in Python
  • integrate the SDK into native C or C++ applications through denkflow.h

Fastest path to first inference​

  1. Export a .denkflow file from the Vision AI Hub.
  2. Request and activate a license.
  3. Install the SDK for your target runtime and language.
  4. Initialize a pipeline from the exported file.
  5. Publish an image, run the pipeline, and read the result.

Continue with Quick Start if you want the shortest path.

Choose your path​

I am upgrading from 0.7​

Read Migrating from 0.7 to 0.8 first; it is the action checklist for breaking changes.

I want the simplest production setup​

Use a .denkflow export and follow:

  1. Quick Start
  2. Installation Guide
  3. Authentication and Licensing
  4. Docker Deployment

I want to assemble pipelines from .denkmodel files​

Use .denkmodel files and follow:

  1. Installation Guide
  2. Runtime and Device Selection
  3. Core Concepts
  4. Creating ImageTensors
  5. Examples

I want to integrate from C or C++​

Follow:

  1. Installation Guide
  2. Quick Start
  3. Creating ImageTensors
  4. Authentication and Licensing

I want to deploy inside containers​

Follow:

  1. Configuration
  2. Docker Deployment
  3. Troubleshooting

Documentation map​