Tag
NVIDIA and Hugging Face publish a hands-on demo showing Gemma 4 running as a vision-language-action model entirely on the Jetson Orin Nano Super, using local STT/TTS and webcam input.
Soul Player C64 implements a real 2-layer decoder-only transformer with ~25,000 int8 parameters in hand-written 6502/6510 assembly, running entirely on an unmodified 1 MHz Commodore 64 loaded from a floppy disk. The project includes training scripts to build and quantize custom models, assemble C64 binaries, and run inference at roughly 60 seconds per token.
Empirical study shows small language models achieve 100% adversarial robustness with System 1 intuition but collapse under System 2 reasoning when used as edge-native governance firewalls in decentralized autonomous organizations.
Cactus-Compute releases Needle, a 26M parameter distilled model from Gemini 3.1, using a pure attention architecture optimized for on-device inference and local fine-tuning.
This blog post details how to set up Frigate with a Hailo AI coprocessor on a Raspberry Pi for object detection, including steps to fix a PCIe descriptor page size error. The setup works with the cheaper Hailo-8L and achieves low inference times.
Google introduces Gemma 3 270M, a compact 270-million parameter model designed for efficient on-device AI with strong instruction-following capabilities and extreme energy efficiency (0.75% battery for 25 conversations on Pixel 9 Pro).
moondream2 is a compact vision language model designed for efficient edge device inference, with benchmark results and usage instructions provided.