Thermal-aware memory layer for LLM agent
Heat-aware memory layer for LLM agents: three-tier hot/warm/cold storage, fusion retrieval (vector cosine + keyword Jaccard + time decay), heat-based eviction with an LLM-distill hook, hot-reload config.
Project overview
Mr. Code Muggle — hi guys, I made something fun to play with: fork it, break it, rebuild it — just maybe mention me (lol). The coral remembers what I can't. Questions? 📮 751286928@qq.com Shoutout to every open-source maker out there 🌱 功能区
蒸馏:相似记忆簇(Jaccard ≥ distillsimthreshold、簇 ≥ distillmincluster)交给 LLM 压缩成 一条 ≤80 字摘要(继承簇的热度/重要性),碎片记忆自动收敛。端点走 OpenAI 兼容 /chat/completions(urllib 零依赖),配置 llm 段(见配置参考)即启用; 未配置或调用失败时优雅降级为"不蒸馏",绝不阻断治理。注意:推理模型(如 deepseek-v4-)的 思考过程也占 maxtokens,本实现已用 1024 保证摘要必出。
治理余量让超容后的淘汰批量发生,而非每次 insert 全量治理(2 万压测:307.8s → 3.4s,90×)。 磁盘配额(storage.maxbytes,0 = 不限制):