Nixtla

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 🚀.

Lightning ⚡️ fast forecasting with statistical and econometric models.

Probabilistic Hierarchical forecasting 👑 with statistical and econometric methods.

Scalable machine 🤖 learning for time series forecasting.

Datasets for time series forecasting

Fast synthetic time series for forecasting, testing, augmentation, and pretraining

Scalable and user friendly neural 🧠 forecasting algorithms.

Helper functions to plot, evaluate, preprocess and engineer features for forecasting

Parser used to render documentation for the `nixtlaverse`