DX-AS (DEEPX All-Suite) is an integrated environment of frameworks and tools that enables inference and compilation of AI models using DEEPX devices. Users can build the integrated environment by installing individual tools, but DX-AS maintains optimal compatibility by aligning the versions of the individual tools.

Purpose

• Must be installed on the Host machine that will perform the compilation (converting) of ONNX models to our proprietary DXNN (DEEPX format).

Core Components

• DX-COM: Converts ONNX models into highly optimized, NPU-ready binaries.

Flexibility & Support

• OS: Compatible with Debian-based Linux (Ubuntu 20.04, 22.04, 24.04)
• Architecture: Supports x86_64 only

Easy Installation

• Our single script automates the full setup process
• All DX-Compiler components are ready to use upon completion.

DX-Compiler provides scripts for local installation, as well as scripts for building Docker images and running containers.

DX-Compiler supports installation in local environments. You can install DX-Compiler by following the instructions at this LINK.

DX-Compiler support installation in docker envirionments. You can install DX-Compiler by following the instructions at this LINK

For detailed instructions on how to run DX-Compiler, please refer to the link below. LINK

To install the necessary Python packages, run the following command:

pip install mkdocs mkdocs-material mkdocs-video pymdown-extensions mkdocs-to-pdf 

To generate the user guide as both HTML and PDF files, execute the following command:

mkdocs build

This will create:

• HTML documentation in the docs/ folder - open docs/index.html in your web browser
• PDF file: DEEPX_DX-COM_UM_[version]_[release_date].pdf in the root directory