#5555th of 56 in ML frameworks and libraries
time-series-library
PyTorch benchmark library of deep time series models for research
- Stars
- 12.9k
- License
- MIT
- Last commit
- Apr 2026
- Language
- Python
Missed the latest GitHub check; these facts may be older.
Overview
TSLib is a PyTorch code base that implements dozens of deep time series models (TimesNet, iTransformer, PatchTST, DLinear, Mamba and others) under one training and evaluation interface. It covers forecasting, imputation, anomaly detection and classification, and supports zero-shot evaluation of pretrained models such as Chronos, TimesFM and Moirai. Maintainers state in April 2026 that no new features will be added.
Who it is for: Deep learning researchers benchmarking time series models
Strengths
- One run.py entry point covers five tasks across many reference model implementations
- Includes zero-shot evaluation for Chronos, TimesFM, Moirai, Sundial and TiRex
- Docker and compose files provided, plus a TimesNet tutorial notebook
- Preprocessed datasets hosted on Google Drive, Baidu and Hugging Face
Weaknesses
- Maintainers will not actively add features; README calls many benchmarks no longer meaningful
- Leaderboard dates from March 2024 and is not current
- Datasets must be downloaded manually into ./dataset
- Mamba needs a Linux, CUDA-specific wheel; torch must match local CUDA
What it needs
- GPU optional
- Docker + Compose
- Needs PyTorch 2.5.1, Conda (Python 3.11), mamba_ssm (for Mamba), uni2ts (for Moirai)
- Models: TimesNet, iTransformer, PatchTST, DLinear, TimeXer
Also in ML frameworks and libraries
See all 56| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 95 out of 100 |
| 2 | pytorchPython tensor library with GPU support and tape-based autograd for deep learning | 94 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 92 out of 100 |
| #4 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 89 out of 100 |
| #5 | pytorch-image-modelsPyTorch library of image models, pretrained weights and reference training scripts | 89 out of 100 |