Show HN: AviGPT-250M – Edge SLM with NVMe memory bus (488MB, runs offline)

AI 导读

AviGPT-250M是首个采用NVMe内存总线的边缘小语言模型，参数量250M，实现100%事实检索与数学精度，效率排名全球第一。

AviGPT-250M is the first edge SLM with NVMe memory bus, achieving 100% factual retrieval and math precision with top efficiency ranking.

重点速览

• 首创NVMe硬件内存总线架构，实现参数与事实存储解耦 Pioneered NVMe memory bus architecture for parameter-memory decoupling
• 事实查询直接调用SSD存储，消除统计猜测误差 Factual queries directly access SSD storage eliminating statistical errors
• 数学计算通过沙箱安全评估器确保100%精度 Math calculations use sandboxed safe math evaluator for 100% accuracy
• 在1.1B参数模型中效率表现超4倍参数量竞品 Outperforms 1.1B-param models 4.4x in efficiency

一句话：边缘计算场景下，硬件加速的SLM可突破传统算力限制。 / Hardware-accelerated SLMs enable edge computing beyond traditional compute limits.

推荐理由 为边缘AI部署提供新范式，展示硬件加速对模型性能的决定性影响。