FiftyOne Agent + Gemma 4

Connect self-hosted open models to the FiftyOne Agent physical AI harness.

Playable

FiftyOne Agent + Gemma 4 media is blocked

Allow external media to connect to the provider and play this content.

Description

An integration of Google's Gemma 4 model served via vLLM into the FiftyOne Agent ecosystem for managing physical AI data. The project demonstrates tool-calling capabilities and memory optimization techniques to run a functional agent brain on a single consumer-grade RTX 4070 Ti GPU.

Descriptions, tags, and model credits may be AI-generated or inferred from public sources and can be incomplete or wrong. Learn more.

Details

Why this is interesting: Demonstrates a tool-calling visual AI agent running on consumer hardware for private physical AI workflows.
Built with: North Mini Code Gemma 4 E2B/E4B
Links: voxel51.com
View tweet
Tags: #physical-ai #computer-vision #self-hosted #vllm

Discovered: X
First posted: X · Aug 2, 2026
Listed: Added 25d ago
Build: Strong
Models: Strong