Edge AI Libraries#
A collection of libraries, microservices, and tools for Edge application development. This project also includes sample applications to showcase some generic AI use cases.
Tools#
Make applications based on sensor data faster, easier, and better.
Visual Pipeline and Platform Evaluation Tool
Computer vision AI models in a fraction of the time and with minimal data.
A robust platform for experimenting with visual prompting techniques.
Libraries#
Media analytics framework for building video AI pipelines with GStreamer.
Open-source anomaly detection library for training and evaluating visual inspection models.
Dataset management framework for preparing, converting, and validating vision datasets.
Motion-control library implementing PLCopen function blocks for industrial automation.
EtherCAT communication stack for deterministic industrial fieldbus integration.
Motion-control task framework for robotics applications and orchestration.
Utilities for segmenting and preprocessing video for downstream AI workflows.
Pluggable Python library for reducing token usage in LLM agent pipelines.
Microservices#
A service that runs an Agentic Predictive Maintenance graph against detections from an external storage API.
A generic, multimodal alert action dispatcher microservice.
Audio analysis microservice for extracting insights from sound inputs.
Containerized service for building and serving video analytics pipelines with DL Streamer.
Document ingestion service using pgvector for embedding storage and retrieval.
A pluggable FastAPI service for routing chat completion requests to multiple inference providers.
Service for collecting and exposing application and system metrics.
Service for fetching, packaging, and preparing models for deployment.
Service for generating and serving embeddings across text and visual inputs.
Service for AI item matching with exact, semantic, or hybrid strategies, returning structured results and confidence scores.
Service for generating speech audio from text input.
Analytics service for ingesting and analyzing time-series data.
Retrieval service for semantic search workflows.
Retrieval service backed by Milvus for semantic search workflows.
Service for preparing multimodal data for retrieval and search workflows.
Service for preparing visual data for retrieval and search workflows.
Service for hierarchical analysis and understanding of video content.
Service for detecting suspicious activity by analyzing video content.
Computes camera parameters automatically.
For object clustering, cluster tracking, and analysis of cluster’s shape and movement patterns.
Combines and contextualizes multiple object detection inputs for tracking objects over time.
Generates meshes and camera parameters from camera-captured frames.
Sample Applications#
Core sample for building retrieval-augmented chat workflows.
Sample application for summarizing documents with LLM-powered workflows.
Sample app for searching video content and generating summaries.
Frameworks#
A set of curated, validated infrastructure profiles, providing a runtime for edge applications.
Model Deployment#
Toolkit for optimizing and deploying AI models on Intel hardware.
Model server for hosting and serving AI models over network APIs.