nixtla
TimeGPT-2.1: production ready pre-trained Time Series Foundation Model for forecasting and anomaly detection. Generative pretrained transformer for time series
What is nixtla?
TimeGPT-2.1: 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 π.
Overview of nixtla
What it is
TimeGPT-2.1 is a production-ready pre-trained Time Series Foundation Model for forecasting and anomaly detection. It is a generative pretrained transformer for time series trained on over 100 billion data points. The model is capable of accurately predicting various domains such as retail, electricity, finance, and IoT with just a few lines of code.
Key features
- Zero-shot Inference: Generate forecasts and detect anomalies without prior training data.
- Fine-tuning: Enhance model performance by fine-tuning on specific datasets.
- API Access: Integrate TimeGPT into applications via a robust API; also deployable on own infrastructure.
- Add Exogenous Variables: Incorporate additional variables (e.g., special dates, events, prices) to improve forecast accuracy.
- Multiple Series Forecasting: Simultaneously forecast multiple time series.
- Prediction Intervals: Provide intervals in predictions to quantify uncertainty.
- Irregular Timestamps: Handle data with non-uniform interval series without preprocessing.
Use cases
TimeGPT can be applied to forecasting and anomaly detection in domains including retail, electricity, finance, IoT, web traffic, and Bitcoin prices. It is suitable for zero-shot inference or fine-tuned on custom data.
Getting started
Install the SDK: python pip install nixtla>=0.8.0 Forecast using TimeGPT in 3 steps (requires an API key from https://nixtla.io/free-trial): python import pandas as pd from nixtla import NixtlaClient nixtla_client = NixtlaClient(api_key = 'YOUR API KEY HERE') df = pd.read_csv('https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/electricity-short.csv') fcst_df = nixtla_client.forecast(df, h=24, level=[80, 90]) nixtla_client.plot(df, fcst_df, level=[80, 90]) For anomaly detection and other languages, see the repository README.
License and pricing
The license is not stated in the repository metadata. The project is listed as open source.
Based on the project's public README and repository metadata.