shiyu-coder/Kronos
Kronos: A Foundation Model for the Language of Financial Markets
Python · ★ 47,304 · Fork 7,819 · 在 GitHub 查看
在手机上浏览源码、订阅每日 AI 趋势日报与项目架构深度解读:获取 CopoHub 客户端(iOS & HarmonyOS)→
GitHub Trending 上榜记录
shiyu-coder/Kronos 已在 GitHub Trending 每日榜单上榜 38 天,最高排名第 2,首次上榜 2026-04-09,最近一次上榜 2026-11-04。
| 日期 | 排名 | Star | 当日增长 |
|---|---|---|---|
| 2026-11-04 | #12 | 46,694 | +503 |
| 2026-11-03 | #8 | 46,490 | +857 |
| 2026-11-02 | #3 | 45,979 | +855 |
| 2026-11-01 | #3 | 45,323 | +611 |
| 2026-10-22 | #9 | 43,044 | +575 |
| 2026-10-21 | #9 | 42,800 | +733 |
| 2026-10-20 | #12 | 42,326 | +719 |
| 2026-10-19 | #12 | 41,611 | +355 |
| 2026-08-13 | #6 | 37,048 | +266 |
| 2026-08-12 | #6 | 36,925 | +277 |
| 2026-08-04 | #10 | 35,935 | +200 |
| 2026-08-03 | #10 | 35,802 | +217 |
| 2026-07-28 | #8 | 34,727 | +441 |
| 2026-07-27 | #7 | 34,535 | +442 |
| 2026-07-26 | #7 | 34,125 | +322 |
| 2026-07-25 | #5 | 33,762 | +499 |
| 2026-07-24 | #3 | 33,463 | +506 |
| 2026-07-23 | #3 | 33,004 | +398 |
| 2026-07-22 | #9 | 32,563 | +134 |
| 2026-06-16 | #18 | 30,448 | +396 |
| 2026-06-15 | #11 | 30,237 | +395 |
| 2026-06-14 | #11 | 29,876 | +238 |
| 2026-05-28 | #8 | 27,066 | +401 |
| 2026-05-27 | #7 | 26,854 | +425 |
| 2026-05-26 | #14 | 26,286 | +245 |
| 2026-05-25 | #10 | 26,003 | +243 |
| 2026-05-24 | #10 | 25,768 | +96 |
| 2026-05-15 | #6 | 24,994 | +363 |
| 2026-05-14 | #6 | 24,775 | +359 |
| 2026-05-07 | #12 | 23,322 | +234 |
| 2026-05-06 | #12 | 23,173 | +241 |
| 2026-04-15 | #9 | 18,022 | +963 |
| 2026-04-14 | #3 | 17,747 | +1,554 |
| 2026-04-13 | #2 | 16,935 | +1,985 |
| 2026-04-12 | #2 | 15,667 | +1,998 |
| 2026-04-11 | #7 | 14,088 | +607 |
| 2026-04-10 | #7 | 12,676 | +602 |
| 2026-04-09 | #9 | 12,116 | +223 |
同期上榜的项目
- NVIDIA/auto-ontology – Auto Ontology is a Python framework for building semantic, graph-based reasoning layers over data. It enables deterministic query generation, better context retrieval, and more explainable AI for LLM and agent workflows.
- max-sixty/worktrunk – Worktrunk is a CLI for Git worktree management, designed for parallel AI agent workflows
- mattpocock/skills – Skills for Real Engineers. Straight from my .agents directory.
- floci-io/floci – Light, fluffy, and always free - The AWS Local Emulator alternative
- thedotmack/claude-mem – Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
- vercel-labs/agent-desktop – Computer use CLI for AI agents
Kronos 项目深度剖析与商业评估报告
Kronos 项目深度剖析与商业评估报告
报告日期: 2026-08-22 | 分析对象: shiyu-coder/Kronos (v1.x) | 分析基准: 37.7k Stars / 6.3k Forks / 264 Open Issues
---
1. 核心定位与商业落地价值
1.1 根本痛点:通用时序模型在金融领域的"语义失配"
Kronos 直击了金融时序建模领域一个长期被忽视的架构级痛点:通用时间序列基础模型(TSFM,如 TimesFM、Chronos、Moirai)以连续数值回归为目标函数,其隐空间表征无法捕捉金融 K 线中具有强语义离散性的微观结构(如跳空缺口、涨跌停、量价背离)。Kronos 通过引入"分层离散 Tokenizer + 自回归 Transformer"的两阶段架构,将金融时序问题显式转化为序列生成问题,从而规避了回归模型在高噪声环境下的过平滑(over-smoothing)问题。
1.2 商业落地场景矩阵
| 场景 | 价值主张 | 目标客户 | |------|----------|----------| | 多资产预测服务 | 45+ 交易所预训练权重覆盖全球主流市场,显著降低冷启动成本 | 量化私募、券商自营 | | 事件驱动风控 | 以生成式分布替代点估计,可量化尾部风险(通过采样分位数) | 银行资管、保险资管 | | 因子工厂 | 微调后的隐藏层输出可替代传统手工因子,作为 ML 因子库输入 | 头部量化对冲基金 | | 智能投顾辅助 | 提供 24h 前瞻概率分布,增强投顾产品的叙事能力 | FinTech 创业公司 |
在 README 中展示上榜徽章
[](https://github.fq6825.top/repo/shiyu-coder/Kronos)