TSFragment-600K
收藏资源简介：
TSFragment-600K是一个包含超过60万个高分辨率时间序列-文本对的片段级数据集，用于文本到时间序列的生成任务。该数据集由香港科技大学（广州）的研究团队创建，包含来自能源消耗、金融汇率、交通、空气质量以及气象变量等多个领域的经典时间序列数据。每个时间序列样本都配有高分辨率、片段级的自然语言描述，通过使用GPT-4o-mini生成，确保了描述的质量和语义丰富性。该数据集为探索文本到时间序列的生成任务提供了坚实的基础，并支持T2S模型进行任意长度时间序列的生成。
TSFragment-600K is a segment-level dataset containing over 600,000 high-resolution time series-text pairs for text-to-time series generation tasks. This dataset was developed by a research team from The Hong Kong University of Science and Technology (Guangzhou), and encompasses classic time series data from multiple domains including energy consumption, financial exchange rates, transportation, air quality, and meteorological variables. Each time series sample is paired with high-resolution, segment-level natural language descriptions generated using GPT-4o-mini, which ensures the quality and semantic richness of the descriptions. This dataset provides a solid foundation for exploring text-to-time series generation tasks, and supports T2S models in generating time series of arbitrary lengths.
数据集概述：TSFragment-600K
基本信息
- 名称：TSFragment-600K
- 发布平台：Hugging Face
- 访问地址：https://huggingface.co/datasets/WinfredGe/TSFragment-600K
- 关联模型：
- T2S-DiT：https://huggingface.co/WinfredGe/T2S-DiT
- T2S-LA-VAE：https://huggingface.co/WinfredGe/T2S-pretrained_LA-VAE
数据集特点
- 规模：600K文本-时间序列片段对
- 领域覆盖：6个经典领域
- 标注类型：细粒度形态描述文本
- 数据对齐：片段级对齐
数据集结构
Data ├─ TSFragment-600K │ ├─ embedding_cleaned_airquality_24.csv │ ├─ embedding_cleaned_airquality_48.csv │ ├─ embedding_cleaned_airquality_96.csv │ ├─ embedding_cleaned_electricity_24.csv │ ├─ embedding_cleaned_electricity_48.csv │ ├─ embedding_cleaned_electricity_96.csv │ │ ... │ ├─ embedding_cleaned_traffic_24.csv │ ├─ embedding_cleaned_traffic_48.csv │ └─ embedding_cleaned_traffic_96.csv ├─ SUSHI │ └─ embedding_cleaned_SUSHI.csv └─ MMD ├─ embedding_cleaned_Agriculture_24.csv ├─ embedding_cleaned_Agriculture_48.csv ├─ embedding_cleaned_Agriculture_96.csv ├─ embedding_cleaned_Climate_24.csv ├─ embedding_cleaned_Climate_48.csv ├─ embedding_cleaned_Climate_96.csv │ ... ├─ embedding_cleaned_SocialGood_24.csv ├─ embedding_cleaned_SocialGood_48.csv └─ embedding_cleaned_SocialGood_96.csv
获取方式
-
直接加载： python from datasets import load_dataset ds = load_dataset("WinfredGe/TSFragment-600K")
-
完整下载：https://drive.google.com/file/d/1tV0xBd0ToWvuLpI5Ocd49uM3QcRkP4NT/view?usp=sharing
应用场景
- 非专家用户通过自然语言描述生成时间序列数据
- 专业人员快速原型设计
- 系统压力测试（模拟极端情况）
相关资源
- 数据集构建管道：./Dataset_Construction_Pipeline/
- 预训练模型：https://drive.google.com/file/d/1T-gjPMvnpSFpkkUSZpAeeIqALThOQydT/view?usp=sharing
引用格式
bibtex @inproceedings{ge2025t2s, title={{T2S}: High-resolution Time Series Generation with Text-to-Series Diffusion Models}, author={Ge, Yunfeng and Li, Jiawei and Zhao, Yiji and Wen, Haomin and Li, Zhao and Qiu, Meikang and Li, Hongyan and Jin, Ming and Pan, Shirui}, booktitle={International Joint Conference on Artificial Intelligence (IJCAI)}, year={2025} }
许可协议
- 许可证类型：Apache-2.0 License
