개요
Kronos는 전 세계 45개 이상의 거래소에서 수집된 금융 K선 데이터로 사전 학습된 디코더 전용 파운데이션 모델 시리즈입니다. 전용 토크나이저와 자기회귀 Transformer를 포함하는 새로운 2단계 프레임워크를 채택하여 예측과 같은 다양한 퀀트 작업을 하나의 통합 모델로 처리할 수 있습니다.
README 미리보기
\r\n Kronos: A Foundation Model for the Language of Financial Markets \r\n\r\n\r\n\r\n\r\n\r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n \r\n\r\n\r\n\r\n\r\n\r\n \r\n Deutsch | \r\n Español | \r\n Français | \r\n 日本語 | \r\n 한국어 | \r\n Português | \r\n Русский | \r\n 中文\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n\r\n> Kronos is the **first open-source foundation model** for financial candlesticks (K-lines), \r\n> trained on data from over **45 global exchanges**.\r\n\r\n\r\n\r\n\r\n## 📰 News\r\n* 🚩 **[2025.11.10]** Kronos has been accpeted by AAAI 2026.\r\n* 🚩 **[2025.08.17]** We have released the scripts for fine-tuning! Check them out to adapt Kronos to your own tasks.\r\n* 🚩 **[2025.08.02]** Our paper is now available on [arXiv](https://arxiv.org/abs/2508.02739)!\r\n\r\n\r\n\r\n## 📜 Introduction\r\n\r\n**Kronos** is a family of decoder-only foundation models, pre-trained specifically for the "language" of financial markets—K-line sequences. Unlike general-purpose TSFMs, Kronos is designed to handle the unique, high-noise characteristics of financial data. It leverages a novel two-stage framework: \r\n1. A specialized tokenizer first quantizes continuous, multi-dimensional K-line data (OHLCV) into **hierarchical discrete tokens**. \r\n2. A large, autoregressive Transformer is then pre-trained on these tokens, enabling it to serve as a unified model for diverse quantitative tasks.\r\n\r\n\r\n \r\n\r\n\r\n## ✨ Live Demo \r\nWe have set up a live demo to visualize Kronos's forecasting results. The webpage showcases a forecast for the **BTC/USDT** trading pair over the next 24 hours. \r\n\r\n**👉 [Access the Live Demo Here](https://shiyu-coder.github.io/Kronos-demo/)** \r\n\r\n## 📦 Model Zoo \r\nWe release a family of pre-trained models with varying capacities to suit different computational and application needs. All models are readily accessible from the Hugging Face Hub.\r\n\r\n| Model | Tokenizer