Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Moondream: Local Vision Model
Learn how Moondream's compact vision-language model runs locally in near real-time, achieving performance comparable to larger VLMs on edge devices.
Moondream is a small vision language model that can easily be run locally, but still performs on par or better than many bigger VLMs.
Moondream: 1GB, 4-bit VLM understands images via text prompts.
Moondream provides a tiny vision-language model for efficient image understanding.
Related projects
Moondream 2 in Rust
Seattle
This talk covers implementing the Moondream 2 vision model in Rust using the Candle ML framework, highlighting Rust…
Day Dream
New York City
Day Dream demonstrates how generative AI can transform physical and virtual environments, enabling real-time exploration of creative possibilities…
vlite 2
Palo Alto
Learn how a lightweight NumPy‑based vector database can replace traditional stores, serving as a fast context library and…
Dream It, Generate It, Animate It: Open-Source Personal Video Creation
Paris
Learn how to combine FLUX1.dev image generation with a text‑to‑video diffusion model, covering pipeline architecture, code walkthrough, scaling,…
Demo - AI Powered, rich video composition in real-time
Bengaluru
See how AI enables real-time, rich video composition for recording and live streaming, including voice anonymity. A mainstage…
Real Time Video AI with ComfyUI
New York City
Learn how to integrate ComfyUI with the ComfyStream framework for real‑time video AI, covering content creation, live communication,…