ONE WARE Documentation

Welcome to the ONE WARE documentation. Find everything you need to build, train, and deploy Vision AI and Edge AI solutions.

Getting Started

New to ONE WARE? Start here:

• What is ONE WARE? — Overview of the ecosystem
• Setup Software — Install ONE WARE Studio, ONE AI Extension, and connect to Cloud
• Your First AI Project — Train and test a model in minutes

Core Documentation

Section	Description
ONE AI	Dataset handling, model configuration, training, export & deployment
Demo Projects	Step-by-step demos for classification, detection, and segmentation
Guides	How-to guides for ONE AI workflows
Cloud & API	REST API reference and self-hosted worker setup

Solutions & Resources

• Benchmarks — Real-world performance comparisons
• Supported Hardware — FPGA vendors, CPU/GPU/TPU targets
• Partners — Technology partners
• Pricing — Credits and licensing

Quick Links

• Quick Start Guide
• Demo Overview
• Cloud API Reference
• Discord Community
• GitHub

ONE AI is a software that generates AI models perfectly fitting your hardware. It takes labeled image data, analyzes task requirements and hardware constraints, and produces deployment-ready models in formats including ONNX, TensorFlow Lite, and VHDL.

Get Started

How do I use ONE AI?

You interact with ONE AI entirely through ONE WARE Studio and ONE WARE Cloud. Studio provides the interface for importing images, labeling data, configuring training settings, and exporting models. Cloud handles job execution and model optimization in the background. You don't need to install or manage ONE AI directly — it runs as part of the platform whenever you start a training or export job.

Documentation Structure

Configuration

Page	Description
Annotation Mode	Setup the annotation mode
Dataset management	General dataset management
Hardware Settings	Target hardware definition and constraint configuration
Prefilters	Image preprocessing pipeline — resolution, crop, color, frequency, threshold filters
Augmentations	Training-time data augmentation — geometric and photometric transforms
Model Settings	Model architecture parameters, prediction types, and resource allocation

Data

Page	Description
Dataset Management	General dataset management and data splits
Import existing data	Import existing images
Capture images from a camera	Use a camera to capture images used for ONE AI

Training

Page	Description
Create a ONE AI model	Create a model for training, test and export
Training	Training configuration and monitoring
Evaluate results	Test and evaluate results
Export	Model export configuration and formats
Integrate	Integrate into your toolchain

Deployment

Page	Description
Model I/O	Input/output specifications for deployed models
FPGA Deployment	VHDL export and FPGA integration
VHDL Documentation	Generated VHDL module reference
C++ API	Native C++ inference API
C# ONNX SDK	.NET ONNX runtime integration

API & Integration

Page	Description
Project File (.oneai)	JSON file format specification for ONE AI projects
REST API	Cloud REST API for full programmatic access
Self hosted training	Self hosted ONE AI Worker

Supported Annotation Modes

Mode	Output	Use Case
Classification	Class labels per image	Defect detection, sorting, counting
Object Detection	Bounding boxes with class labels	Locating and identifying multiple objects
Segmentation	Per-pixel class masks	Precise region delineation

Supported Export Formats

Format	Target
ONNX	CPU, GPU, edge devices via ONNX Runtime
TensorFlow Lite	Microcontrollers, mobile, TPU
VHDL	FPGA synthesis
Binary / Source	Standalone executables, embedded integration

Image Fusion Modes

Mode	Input	Description
Single	1 image	Standard single-frame inference
Multi	N images	Multiple viewpoints of the same scene
Difference	2 images	Detects changes between reference and current frame
Comparison	2 images	Compares two arbitrary images
Stereo	2 images	Left/right stereo pair for depth-aware detection

Need Help? We're Here for You!

Christopher from our development team is ready to help with any questions about ONE AI usage, troubleshooting, or optimization. Don't hesitate to reach out!

Our Support Email:support@one-ware.com