[GitHub Trending] cactus-compute/needle
Scored daily by a customisable AI persona to surface the most relevant engineering leadership news.
Tiny foundation model for edge devices, highly novel and relevant to AI/ML deployment trends.
Needle 2 is a 45M-parameter open model for tool calling and structured extraction, compressed into a single 14MB binary that runs a full session in ~28MB RAM. Built on a Simple Attention Network with Hadamard MLP and CQ2-bit quantization, it trades wins with models 5-70x larger like FunctionGemma 270M. The Python package provides inference, LoRA fine-tuning, and export; tool calls return structured JSON constrained by a byte-level grammar, with confidence-gated responses and built-in retrieval that selects the top five tools per turn from a large catalogue.
cactus-compute