DevTrends开发者趋势
主题色
HelloGitHub 投稿 ↗

NineNi999neNine /

FractalBrainOS

Project URL Category C++ Project Title An open-source, non-backprop neuromorphic brain engine built in C++17 using Kuramoto oscillators and Free Energy Principle. Project FractalBrainOS is a self-learning hierarchical os…

原文
DT研究与学习DT数据与运维语言C++原始active-inference原始agi原始artificial-general-intelligence原始bio-inspired-ai原始brain-simulation原始complex-systems原始cpp17原始decentralized-ai原始edge-ai

项目介绍

暂无此语言译文,以下保留原文。

Project URL Category C++ Project Title An open-source, non-backprop neuromorphic brain engine built in C++17 using Kuramoto oscillators and Free Energy Principle. Project FractalBrainOS is a self-learning hierarchical oscillatory neural network. It runs continuously on edge hardware (Raspberry Pi/ARM) using int16 quantization, minimizes free energy without backpropagation, and uses LLMs as a speech zone. Highlights No Backprop / GPU: Learns dynamically via Kuramoto phase synchronization and Friston's Free Energy Principle. Ultra-efficient: int16 quantization and precomputed sin/cos LUT tables (512KB footprint) optimized for edge computing on ARM/Raspberry Pi. Multi-scale Memory: Clock oscillators (0.1s to 24h) providing temporal addressing alongside standing-wave delay lines. Ethical Core: Hardcoded Kubera Dharma protocol (Ahimsa) integrated directly into dynamical math. Example Code markdown Compile g++ -O3 -std=c++17 -pthread -o brain brainv5.cpp Run test simulation (Level 6) ./brain --test 6 or Videos

介绍整理自已收录的社区资料。

收录记录

以下日期为 DevTrends 收录日期,指标来自当时的数据快照。

  1. Project URL Category C++ Project Title An open-source, non-backprop neuromorphic brain engine built in C++17 using Kuramoto oscillators and Free Energy Principle. Project FractalBrainOS is a self-learning hierarchical oscillatory neural network. It runs continuously on edge hardware (Raspberry Pi/ARM) using int16 quantization, minimizes free energy without backpropagation, and uses LLMs as a speech zone. Highlights No Backprop / GPU: Learns dynamically via Kuramoto phase synchronization and Friston's Free Energy Principle. Ultra-efficient: int16 quantization and precomputed sin/cos LUT tables (512KB footprint) optimized for edge computing on ARM/Raspberry Pi. Multi-scale Memory: Clock oscillators (0.1s to 24h) providing temporal addressing alongside standing-wave delay lines. Ethical Core: Hardcoded Kubera Dharma protocol (Ahimsa) integrated directly into dynamical math. Example Code markdown Compile g++ -O3 -std=c++17 -pthread -o brain brainv5.cpp Run test simulation (Level 6) ./brain --test 6 or Videos

    原文

DEV TRENDS / COMMUNITY

讨论这个项目

使用 GitHub 登录,分享使用体验、问题或项目新动态。

查看全部讨论 ↗