flowzl/timelib
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
main
timelib/README_zh.md
flowerstonezl f7291b1221 first commit
2026-01-29 17:08:54 +08:00

330 lines
26 KiB
Markdown
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# 时间序列库（TSLib）
TSLib 是一个面向深度学习研究者的开源库，特别适用于深度时间序列分析。
> **English README**：[README.md](./README.md)
我们提供了一个整洁的代码库，用于评测先进的深度时间序列模型或开发自定义模型，覆盖 **长短期预测、插补、异常检测和分类** 等五大主流任务。
:triangular_flag_on_post:**最新动态**（2025.11）鉴于大型时间序列模型（LTSM）的快速发展，我们在 TSLib 中新增了[[零样本预测]](https://github.com/thuml/Time-Series-Library/blob/main/exp/exp_zero_shot_forecasting.py)功能，可参考 [此脚本](https://github.com/thuml/Time-Series-Library/blob/main/scripts/long_term_forecast/ETT_script/LTSM.sh) 评测 LTSM。
:triangular_flag_on_post:**最新动态**（2025.10）针对近期研究者在标准基准上追求微小提升而产生的困惑，我们提出了[[精度定律]](https://arxiv.org/abs/2510.02729)，以刻画深度时间序列预测任务的目标，并可据此识别已饱和的数据集。
:triangular_flag_on_post:**最新动态**（2024.10）我们已纳入 [[TimeXer]](https://arxiv.org/abs/2402.19072)，其定义了一个实用的预测范式：带外生变量的预测。考虑到实用性与计算效率，我们认为 TimeXer 所定义的新范式将成为未来研究的“正确”任务。
:triangular_flag_on_post:**最新动态**（2024.10）实验室已开源 [[OpenLTM]](https://github.com/thuml/OpenLTM)，提供了有别于 TSLib 的预训练 - 微调范式。如果您对大型时间序列模型感兴趣，该仓库值得参考。
:triangular_flag_on_post:**最新动态**（2024.07）我们撰写了关于[[深度时间序列模型]](https://arxiv.org/abs/2407.13278)的综述，并基于 TSLib 构建了严谨的基准。论文总结了当前时间序列模型的设计原则，并通过深入实验验证，期望对未来研究有所帮助。
:triangular_flag_on_post:**最新动态**（2024.04）感谢 [frecklebars](https://github.com/thuml/Time-Series-Library/pull/378) 的贡献，著名的序列模型 [Mamba](https://arxiv.org/abs/2312.00752) 已加入本库。参见[该文件](https://github.com/thuml/Time-Series-Library/blob/main/models/Mamba.py)，需要先用 pip 安装 `mamba_ssm`。
:triangular_flag_on_post:**最新动态**（2024.03）鉴于各论文使用的回溯窗口长度不一致，我们将排行榜中的长期预测拆分为 Look-Back-96 与 Look-Back-Searching 两类。建议阅读 [TimeMixer](https://openreview.net/pdf?id=7oLshfEIC2)，其实验同时包含两种窗口设置，更具科学性。
:triangular_flag_on_post:**最新动态**（2023.10）我们添加了 [iTransformer](https://arxiv.org/abs/2310.06625) 的实现，这是长期预测领域的最新 SOTA。官方代码与完整脚本参见 [此处](https://github.com/thuml/iTransformer)。
:triangular_flag_on_post:**最新动态**（2023.09）我们为 [TimesNet](https://openreview.net/pdf?id=ju_Uqw384Oq) 及本库添加了详细[教程](https://github.com/thuml/Time-Series-Library/blob/main/tutorial/TimesNet_tutorial.ipynb)，对时间序列初学者十分友好。
:triangular_flag_on_post:**最新动态**（2023.02）我们发布了 TSlib，作为一个面向时间序列模型的综合基准与代码库，扩展自此前的 [Autoformer](https://github.com/thuml/Autoformer) 仓库。
## 时间序列分析排行榜
截至 2024 年 3 月，各任务排行榜前三名如下：
| 模型<br>排名 | 长期预测<br>Look-Back-96 | 长期预测<br/>Look-Back-Searching | 短期预测 | 插补 | 分类 | 异常检测 |
| ------------ | ------------------------ | -------------------------------- | -------- | ---- | ---- | -------- |
| 🥇 第一名 | [TimeXer](https://arxiv.org/abs/2402.19072) | [TimeMixer](https://openreview.net/pdf?id=7oLshfEIC2) | [TimesNet](https://arxiv.org/abs/2210.02186) | [TimesNet](https://arxiv.org/abs/2210.02186) | [TimesNet](https://arxiv.org/abs/2210.02186) | [TimesNet](https://arxiv.org/abs/2210.02186) |
| 🥈 第二名 | [iTransformer](https://arxiv.org/abs/2310.06625) | [PatchTST](https://github.com/yuqinie98/PatchTST) | [Non-stationary<br/>Transformer](https://github.com/thuml/Nonstationary_Transformers) | [Non-stationary<br/>Transformer](https://github.com/thuml/Nonstationary_Transformers) | [Non-stationary<br/>Transformer](https://github.com/thuml/Nonstationary_Transformers) | [FEDformer](https://github.com/MAZiqing/FEDformer) |
| 🥉 第三名 | [TimeMixer](https://openreview.net/pdf?id=7oLshfEIC2) | [DLinear](https://arxiv.org/pdf/2205.13504.pdf) | [FEDformer](https://github.com/MAZiqing/FEDformer) | [Autoformer](https://github.com/thuml/Autoformer) | [Informer](https://github.com/zhouhaoyi/Informer2020) | [Autoformer](https://github.com/thuml/Autoformer) |
**说明：排行榜会持续更新。** 如果您提出了先进的模型，可通过发送论文或代码链接、或提交 PR 与我们联系，我们会尽快将其加入仓库并更新排行榜。
**排行榜中的对比模型**（☑ 表示代码已收录）。
- [x] **TimeXer** - TimeXer: Empowering Transformers for Time Series Forecasting with Exogenous Variables [[NeurIPS 2024]](https://arxiv.org/abs/2402.19072) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TimeXer.py)
- [x] **TimeMixer** - TimeMixer: Decomposable Multiscale Mixing for Time Series Forecasting [[ICLR 2024]](https://openreview.net/pdf?id=7oLshfEIC2) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TimeMixer.py)
- [x] **TSMixer** - TSMixer: An All-MLP Architecture for Time Series Forecasting [[arXiv 2023]](https://arxiv.org/pdf/2303.06053.pdf) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TSMixer.py)
- [x] **iTransformer** - iTransformer: Inverted Transformers Are Effective for Time Series Forecasting [[ICLR 2024]](https://arxiv.org/abs/2310.06625) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/iTransformer.py)
- [x] **PatchTST** - A Time Series is Worth 64 Words: Long-term Forecasting with Transformers [[ICLR 2023]](https://openreview.net/pdf?id=Jbdc0vTOcol) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/PatchTST.py)
- [x] **TimesNet** - TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis [[ICLR 2023]](https://openreview.net/pdf?id=ju_Uqw384Oq) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TimesNet.py)
- [x] **DLinear** - Are Transformers Effective for Time Series Forecasting? [[AAAI 2023]](https://arxiv.org/pdf/2205.13504.pdf) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/DLinear.py)
- [x] **LightTS** - Less Is More: Fast Multivariate Time Series Forecasting with Light Sampling-oriented MLP Structures [[arXiv 2022]](https://arxiv.org/abs/2207.01186) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/LightTS.py)
- [x] **ETSformer** - ETSformer: Exponential Smoothing Transformers for Time-series Forecasting [[arXiv 2022]](https://arxiv.org/abs/2202.01381) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/ETSformer.py)
- [x] **Non-stationary Transformer** - Non-stationary Transformers: Exploring the Stationarity in Time Series Forecasting [[NeurIPS 2022]](https://openreview.net/pdf?id=ucNDIDRNjjv) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Nonstationary_Transformer.py)
- [x] **FEDformer** - FEDformer: Frequency Enhanced Decomposed Transformer for Long-term Series Forecasting [[ICML 2022]](https://proceedings.mlr.press/v162/zhou22g.html) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/FEDformer.py)
- [x] **Pyraformer** - Pyraformer: Low-complexity Pyramidal Attention for Long-range Time Series Modeling and Forecasting [[ICLR 2022]](https://openreview.net/pdf?id=0EXmFzUn5I) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Pyraformer.py)
- [x] **Autoformer** - Autoformer: Decomposition Transformers with Auto-Correlation for Long-Term Series Forecasting [[NeurIPS 2021]](https://openreview.net/pdf?id=I55UqU-M11y) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Autoformer.py)
- [x] **Informer** - Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting [[AAAI 2021]](https://ojs.aaai.org/index.php/AAAI/article/view/17325/17132) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Informer.py)
- [x] **Reformer** - Reformer: The Efficient Transformer [[ICLR 2020]](https://openreview.net/forum?id=rkgNKkHtvB) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Reformer.py)
- [x] **Transformer** - Attention is All You Need [[NeurIPS 2017]](https://proceedings.neurips.cc/paper/2017/file/3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Transformer.py)
更多详情可参考我们关于 [[TimesNet]](https://arxiv.org/abs/2210.02186) 的最新论文，实时在线版本即将发布。
**新增基线模型**（综合评测后将加入排行榜）。
- [x] **TimeFilter** - TimeFilter: Patch-Specific Spatial-Temporal Graph Filtration for Time Series Forecasting [[ICML 2025]](https://arxiv.org/abs/2501.13041) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TimeFilter.py)
- [x] **KAN-AD** - KAN-AD: Time Series Anomaly Detection with Kolmogorov-Arnold Networks [[ICML 2025]](https://arxiv.org/abs/2411.00278) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/KANAD.py)
- [x] **MultiPatchFormer** - A multiscale model for multivariate time series forecasting [[Scientific Reports 2025]](https://www.nature.com/articles/s41598-024-82417-4) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/MultiPatchFormer.py)
- [x] **WPMixer** - WPMixer: Efficient Multi-Resolution Mixing for Long-Term Time Series Forecasting [[AAAI 2025]](https://arxiv.org/abs/2412.17176) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/WPMixer.py)
- [x] **MSGNet** - MSGNet: Learning Multi-Scale Inter-Series Correlations for Multivariate Time Series Forecasting [[AAAI 2024]](https://dl.acm.org/doi/10.1609/aaai.v38i10.28991) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/MSGNet.py)
- [x] **PAttn** - Are Language Models Actually Useful for Time Series Forecasting? [[NeurIPS 2024]](https://arxiv.org/pdf/2406.16964) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/PAttn.py)
- [x] **Mamba** - Mamba: Linear-Time Sequence Modeling with Selective State Spaces [[arXiv 2023]](https://arxiv.org/abs/2312.00752) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Mamba.py)
- [x] **SegRNN** - SegRNN: Segment Recurrent Neural Network for Long-Term Time Series Forecasting [[arXiv 2023]](https://arxiv.org/abs/2308.11200.pdf) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/SegRNN.py)
- [x] **Koopa** - Koopa: Learning Non-stationary Time Series Dynamics with Koopman Predictors [[NeurIPS 2023]](https://arxiv.org/pdf/2305.18803.pdf) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Koopa.py)
- [x] **FreTS** - Frequency-domain MLPs are More Effective Learners in Time Series Forecasting [[NeurIPS 2023]](https://arxiv.org/pdf/2311.06184.pdf) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/FreTS.py)
- [x] **MICN** - MICN: Multi-scale Local and Global Context Modeling for Long-term Series Forecasting [[ICLR 2023]](https://openreview.net/pdf?id=zt53IDUR1U) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/MICN.py)
- [x] **Crossformer** - Crossformer: Transformer Utilizing Cross-Dimension Dependency for Multivariate Time Series Forecasting [[ICLR 2023]](https://openreview.net/pdf?id=vSVLM2j9eie) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Crossformer.py)
- [x] **TiDE** - Long-term Forecasting with TiDE: Time-series Dense Encoder [[arXiv 2023]](https://arxiv.org/pdf/2304.08424.pdf) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TiDE.py)
- [x] **SCINet** - SCINet: Time Series Modeling and Forecasting with Sample Convolution and Interaction [[NeurIPS 2022]](https://openreview.net/pdf?id=AyajSjTAzmg) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/SCINet.py)
- [x] **FiLM** - FiLM: Frequency improved Legendre Memory Model for Long-term Time Series Forecasting [[NeurIPS 2022]](https://openreview.net/forum?id=zTQdHSQUQWc) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/FiLM.py)
- [x] **TFT** - Temporal Fusion Transformers for Interpretable Multi-horizon Time Series Forecasting [[arXiv 2019]](https://arxiv.org/abs/1912.09363) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TemporalFusionTransformer.py)
**新增大型时间序列模型**。本库同样支持以下 LTSM 的零样本评测：
- [x] **Chronos2** - Chronos-2: From Univariate to Universal Forecasting [[arXiv 2025]](https://arxiv.org/abs/2510.15821) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Chronos2.py)
- [x] **TiRex** - TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learning [[NeurIPS 2025]](https://arxiv.org/pdf/2505.23719) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TiRex.py)
- [x] **Sundial** - Sundial: A Family of Highly Capable Time Series Foundation Models [[ICML 2025]](https://arxiv.org/pdf/2502.00816) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Sundial.py)
- [x] **Time-MoE** - Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts [[ICLR 2025]](https://arxiv.org/pdf/2409.16040) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TimeMoE.py)
- [x] **Toto** - Toto: Time Series Optimized Transformer for Observability [[arXiv 2024]](https://arxiv.org/pdf/2407.07874)
- [x] **Chronos** - Chronos: Learning the Language of Time Series [[TMLR 2024]](https://arxiv.org/pdf/2403.07815) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/Chronos.py)
- [x] **Moirai** - Unified Training of Universal Time Series Forecasting Transformers [[ICML 2024]](https://arxiv.org/pdf/2402.02592)
- [x] **TimesFM** - TimesFM: A decoder-only foundation model for time-series forecasting [[ICML 2024]](https://arxiv.org/abs/2310.10688) [[代码]](https://github.com/thuml/Time-Series-Library/blob/main/models/TimesFM.py)
## 快速开始
### 准备数据
可从 [[Google Drive]](https://drive.google.com/drive/folders/13Cg1KYOlzM5C7K8gK8NfC-F3EYxkM3D2?usp=sharing)、[[Baidu Drive]](https://pan.baidu.com/s/1r3KhGd0Q9PJIUZdfEYoymg?pwd=i9iy) 或 [[Hugging Face]](https://huggingface.co/datasets/thuml/Time-Series-Library) 下载预处理数据，并置于 `./dataset` 目录。
### 安装
1. 克隆本仓库
```bash
git clone https://github.com/thuml/Time-Series-Library.git
cd Time-Series-Library
```
2. 创建新的 Conda 环境
```bash
conda create -n tslib python=3.11
conda activate tslib
```
3. 安装核心依赖
```bash
pip install -r requirements.txt
```
4. 安装 Mamba 模型依赖（models/Mamba.py 需要）
> ⚠️ **CUDA 兼容性提示**
> Mamba 预编译包与 **CUDA 版本强相关**。
> 请确保安装与本地 CUDA 版本匹配的包（如 `cu11` 或 `cu12`）。
> 版本不匹配可能导致运行时错误或导入失败。
**CUDA 12** 示例：
```bash
pip install https://github.com/state-spaces/mamba/releases/download/v2.2.6.post3/mamba_ssm-2.2.6.post3+cu12torch2.5cxx11abiFALSE-cp311-cp311-linux_x86_64.whl
```
5. 安装 Moirai 模型依赖（models/Moirai.py 需要）
```bash
pip install uni2ts --no-deps
```
### Docker 部署
```bash
# 构建并以后台模式启动容器
docker compose -f 'Time-Series-Library/docker-compose.yml' up -d --build
# 在仓库根目录创建 ./dataset 并下载/放置数据集
mkdir -p dataset
# 将本地数据集复制到容器内 /workspace/dataset
docker cp ./dataset tslib:/workspace/dataset
# 进入运行中的容器
docker exec -it tslib bash
# 切换到容器内的工作目录
cd /workspace
# 使用预训练 Moirai 模型进行零样本预测
python -u run.py \
--task_name zero_shot_forecast \ # 任务类型：零样本预测
--is_training 0 \ # 0 = 仅推理
--root_path ./dataset/ETT-small/ \ # 数据集根路径
--data_path ETTh1.csv \ # 数据文件名
--model_id ETTh1_512_96 \ # 实验/模型标识
--model Moirai \ # 模型名称（TimesFM / Moirai）
--data ETTh1 \ # 数据集名称
--features M \ # 多变量预测
--seq_len 512 \ # 输入序列长度
--pred_len 96 \ # 预测步长
--enc_in 7 \ # 输入变量数
--des 'Exp' \ # 实验描述
--itr 1 # 运行次数
```
### 快速测试
5个任务快速测试（每个任务1个epoch）：
```bash
# 执行所有5个任务的快速测试
export CUDA_VISIBLE_DEVICES=0
# 1. 长期预测
python -u run.py --task_name long_term_forecast --is_training 1 --root_path ./dataset/ETT-small/ --data_path ETTh1.csv --model_id test_long --model DLinear --data ETTh1 --features M --seq_len 96 --pred_len 96 --enc_in 7 --dec_in 7 --c_out 7 --train_epochs 1 --num_workers 2
# 2. 短期预测（使用ETT数据集，较短预测长度）
python -u run.py --task_name long_term_forecast --is_training 1 --root_path ./dataset/ETT-small/ --data_path ETTh1.csv --model_id test_short --model TimesNet --data ETTh1 --features M --seq_len 24 --label_len 12 --pred_len 24 --e_layers 2 --d_layers 1 --d_model 16 --d_ff 32 --enc_in 7 --dec_in 7 --c_out 7 --top_k 5 --train_epochs 1 --num_workers 2
# 3. 插补
python -u run.py --task_name imputation --is_training 1 --root_path ./dataset/ETT-small/ --data_path ETTh1.csv --model_id test_imp --model TimesNet --data ETTh1 --features M --seq_len 96 --e_layers 2 --d_layers 1 --d_model 16 --d_ff 32 --enc_in 7 --dec_in 7 --c_out 7 --top_k 3 --train_epochs 1 --num_workers 2 --label_len 0 --pred_len 0 --mask_rate 0.125 --learning_rate 0.001
# 4. 异常检测
python -u run.py --task_name anomaly_detection --is_training 1 --root_path ./dataset/PSM --model_id test_ad --model TimesNet --data PSM --features M --seq_len 100 --pred_len 0 --d_model 64 --d_ff 64 --e_layers 2 --enc_in 25 --c_out 25 --anomaly_ratio 1.0 --top_k 3 --train_epochs 1 --batch_size 128 --num_workers 2
# 5. 分类
python -u run.py --task_name classification --is_training 1 --root_path ./dataset/Heartbeat/ --model_id Heartbeat --model TimesNet --data UEA --e_layers 2 --d_layers 1 --factor 3 --d_model 64 --d_ff 128 --top_k 3 --train_epochs 1 --batch_size 16 --learning_rate 0.001 --num_workers 0
```
### 训练与评测
`./scripts/` 目录下提供了全部基准的实验脚本，可参考下列示例复现实验：
```bash
# 长期预测
bash ./scripts/long_term_forecast/ETT_script/TimesNet_ETTh1.sh
# 短期预测
bash ./scripts/short_term_forecast/TimesNet_M4.sh
# 插补
bash ./scripts/imputation/ETT_script/TimesNet_ETTh1.sh
# 异常检测
bash ./scripts/anomaly_detection/PSM/TimesNet.sh
# 分类
bash ./scripts/classification/TimesNet.sh
```
### 开发自定义模型
- 将模型文件放入 `./models`，可参考 `./models/Transformer.py`。
- 在 `./exp/exp_basic.py` 的 `Exp_Basic.model_dict` 中注册新模型。
- 在 `./scripts` 下创建对应的运行脚本。
### 注意事项：
(1) 关于分类：由于我们在统一代码库中涵盖五大任务，各子任务的精度可能略有波动，但平均性能可复现（甚至略高）。复现用 checkpoint 可在 [此处](https://github.com/thuml/Time-Series-Library/issues/494) 下载。
(2) 关于异常检测：有关异常检测调整策略的讨论见[这里](https://github.com/thuml/Anomaly-Transformer/issues/14)，核心是该调整策略对应事件级指标。
### 查看项目文件结构：
```
Time-Series-Library/
├── README.md # 官方README，包含任务、榜单、使用方法
├── requirements.txt # pip依赖列表，直接pip install复现环境
├── LICENSE / CONTRIBUTING.md # 原项目许可与贡献指南
├── run.py # 单入口脚本，解析参数并调度各任务
├── exp/ # 各任务实验管线，封装训练/验证/测试
│ ├── exp_basic.py # 实验基类，注册所有模型，统一构建流程
│ ├── exp_long_term_forecasting.py # 长期预测实验逻辑
│ ├── exp_short_term_forecasting.py # 短期预测实验逻辑
│ ├── exp_imputation.py # 缺失值填充实验
│ ├── exp_anomaly_detection.py # 异常检测实验
│ ├── exp_classification.py # 分类实验
│ └── exp_zero_shot_forecasting.py # LTSM零样本预测评估
├── data_provider/ # 数据入口，负责数据集载入与切分
│ ├── data_factory.py # 根据任务选择对应DataLoader
│ ├── data_loader.py # 通用时序数据读取与滑窗逻辑
│ ├── uea.py / m4.py # UEA、M4等特定数据格式处理
│ └── __init__.py # 暴露上层可用的数据工厂接口
├── models/ # 所有模型实现，文件名即模型名
│ ├── TimesNet.py、TimeMixer.py 等 # 主流预测模型
│ ├── Chronos2.py、TiRex.py # LTSM零样本模型
│ └── __init__.py # 统一导出供实验模块按名称实例化
├── layers/ # 复用层/块，如注意力、卷积、嵌入
│ ├── Transformer_EncDec.py # Transformer编解码堆栈
│ ├── AutoCorrelation.py # 自相关算子
│ ├── MultiWaveletCorrelation.py# 频域单元
│ └── Embed.py 等 # 各模型共享基元
├── utils/ # 工具集合
│ ├── metrics.py # MSE/MAE/DTW等评估指标
│ ├── tools.py # 训练通用工具，比如EarlyStopping
│ ├── augmentation.py # 分类/检测任务增强策略
│ ├── print_args.py # 统一打印参数
│ └── masking.py / losses.py # 任务相关辅助函数
├── scripts/ # 复现实验的bash脚本
│ ├── long_term_forecast/ # 按数据集/模型划分的长期预测脚本
│ ├── short_term_forecast/ # M4等短期预测脚本
│ ├── imputation/ # 多数据集缺失填充脚本
│ ├── anomaly_detection/ # SMD/SMAP/SWAT等检测脚本
│ ├── classification/ # UEA分类脚本
│ └── exogenous_forecast/ # TimeXer外生变量预测流程
├── tutorial/ # 官方TimesNet教学notebook与插图
└── pic/ # README插图（数据集分布等）
```
### 理解项目架构：
- **整体流程**：通过 `scripts/*.sh` 设定实验参数 → 调用 `python run.py ...` → `run.py` 解析参数并根据 `task_name` 选择对应 `Exp_*` 类 → `Exp_*` 内部利用 `data_provider` 构造数据加载器、`models` 实例化网络、`utils` 中的工具完成训练/验证/测试 → 结果与模型参数写入 `./checkpoints`。
- **实验层（exp/）**：`Exp_Basic` 负责注册模型与设备，子类实现 `_get_data/train/test`，将不同任务的差异隔离，方便模型在多任务间复用。
- **模型与层（models/ + layers/）**：模型文件集中定义各网络结构，公用的注意力、卷积、频域块等沉淀在 `layers/`，减少重复实现。
- **数据层（data_provider/）**：`data_factory` 按任务返回 Dataset/DataLoader，`data_loader` 封装序列裁剪、滑动窗口、掩码策略，不同任务通过参数控制窗口长度、缺失率、异常比例。
- **脚本层（scripts/）**：提供与论文一致的复现实验脚本，涵盖各种数据集/模型/GPU 配置，便于批量跑榜，也可作为自定义实验的起点。
- **辅助层（utils/）**：`metrics` 统一评估指标，`tools` 中的 `EarlyStopping`、`adjust_learning_rate` 等负责训练调度；`augmentation`/`masking` 等用于任务特定的数据增强或预处理。
- **学习建议**：阅读顺序推荐 `scripts -> run.py -> exp/exp_basic.py -> 对应 Exp 子类 -> data_provider -> models`，并结合 `tutorial/TimesNet_tutorial.ipynb` 快速熟悉整体调用链，再按需深入模型或层级实现。
## 引用
如果本仓库对您有帮助，请引用以下论文：
```
@inproceedings{wu2023timesnet,
title={TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis},
author={Haixu Wu and Tengge Hu and Yong Liu and Hang Zhou and Jianmin Wang and Mingsheng Long},
booktitle={International Conference on Learning Representations},
year={2023},
}
@article{wang2024tssurvey,
title={Deep Time Series Models: A Comprehensive Survey and Benchmark},
author={Yuxuan Wang and Haixu Wu and Jiaxiang Dong and Yong Liu and Mingsheng Long and Jianmin Wang},
booktitle={arXiv preprint arXiv:2407.13278},
year={2024},
}
```
## 联系方式
如有问题或建议，欢迎联系维护团队：
现任：
- Haixu Wu（博士生，wuhx23@mails.tsinghua.edu.cn）
- Yuxuan Wang（博士生，wangyuxu22@mails.tsinghua.edu.cn）
- Yong Liu（博士生，liuyong21@mails.tsinghua.edu.cn）
- Huikun Weng（本科生，wenghk22@mails.tsinghua.edu.cn）
往届：
- Tengge Hu（硕士，htg21@mails.tsinghua.edu.cn）
- Haoran Zhang（硕士，z-hr20@mails.tsinghua.edu.cn）
- Jiawei Guo（本科生，guo-jw21@mails.tsinghua.edu.cn）
也欢迎在 Issues 中反馈。
## 致谢
本库参考了以下仓库：
- 预测：https://github.com/thuml/Autoformer
- 异常检测：https://github.com/thuml/Anomaly-Transformer
- 分类：https://github.com/thuml/Flowformer
实验所用数据集均为公开数据，来源如下：
- 长期预测与插补：https://github.com/thuml/Autoformer
- 短期预测：https://github.com/ServiceNow/N-BEATS
- 异常检测：https://github.com/thuml/Anomaly-Transformer
- 分类：https://www.timeseriesclassification.com/
## 感谢所有贡献者
<a href="https://github.com/thuml/Time-Series-Library/graphs/contributors">
<img src="https://contrib.rocks/image?repo=thuml/Time-Series-Library" />
</a>
Reference in New Issue View Git Blame Copy Permalink