TimeGPT-1

Azul Garza, Cristian Challu, Max Mergenthaler-Canseco
10/5/2023
cs.LGstat.AP

Official code: Nixtla/nixtla (3,979 stars, last commit 2026-05-01).

Abstract

In this paper, we introduce TimeGPT, the first foundation model for time series, capable of generating accurate predictions for diverse datasets not seen during training. We evaluate our pre-trained model against established statistical, machine learning, and deep learning methods, demonstrating that TimeGPT zero-shot inference excels in performance, efficiency, and simplicity. Our study provides compelling evidence that insights from other domains of artificial intelligence can be effectively applied to time series analysis. We conclude that large-scale time series models offer an exciting opportunity to democratize access to precise predictions and reduce uncertainty by leveraging the capabilities of contemporary advancements in deep learning.

View on arXivView PDF

Code Implementations(1)

Nixtla/nixtlaOfficial68%

TimeGPT-1: production ready pre-trained Time Series Foundation Model for forecasting and anomaly detection. Generative pretrained transformer for time series trained on over 100B data points. It's capable of accurately predicting various domains such as retail, electricity, finance, and IoT with just a few lines of code 🚀.

3,979334R, CSSSep 23, 20215 months agoNOASSERTION
agentagentic-aianomaly-detectionartificial-intelligencedeep-learning+10 more

Cite this paper

@article{garza2023timegpt,
  title  = {TimeGPT-1},
  author = {Azul Garza and Cristian Challu and Max Mergenthaler-Canseco},
  year   = {2023},
  eprint = {2310.03589},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2310.03589v3}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2310.03589.svg)](https://codeofpaper.com/paper/2310.03589)

Discussion