About TimeGPT
Learn about TimeGPT - the foundation model for time series.
TimeGPT is a production-ready generative pretrained transformer model specifically designed for time series forecasting. It accurately forecasts domains such as retail, electricity, finance, and IoT with minimal code. Below you’ll find a high-level overview of its features, architecture, and practical examples.
1. Start Now
2. Choose Your Interface
You can access TimeGPT through:
-
Public APIs
-
Private deployment on your own infrastructure
3. Upload Time Series Data
Provide your time series data. TimeGPT can consume data with irregular timestamps, exogenous variables, or multiple time series simultaneously.
4. Forecast or Detect Anomalies
Perform zero-shot inference out-of-the-box to forecast future values or detect anomalies. Fine-tune the model if you need more targeted performance.
For detailed instructions and advanced configurations, visit our Quickstart Guide and additional tutorials.
Features and Capabilities
Zero-shot Inference
Zero-shot Inference
Zero-shot Inference: Generate forecasts and detect anomalies immediately without prior training. Quickly gain insights from your data.
Fine-tuning
Fine-tuning
Fine-tuning: Enhance prediction accuracy by training TimeGPT on your own datasets, tailoring it to your unique scenario.
API Access
API Access
API Access: Integrate forecasts into applications via a robust API. Easily obtain keys at the Dashboard. Also accessible in Azure Studio or through private deployment.
Add Exogenous Variables
Add Exogenous Variables
Add Exogenous Variables: Incorporate external variables (e.g., events, prices) to improve forecast accuracy.
Multiple Series Forecasting
Multiple Series Forecasting
Multiple Series Forecasting: Predict multiple time series at once, improving workflow efficiency.
Specific Loss Function
Specific Loss Function
Specific Loss Function: Customize training with loss functions that match your performance objectives.
Cross-validation
Cross-validation
Cross-validation: Evaluate model reliability and generalization with built-in cross-validation.
Prediction Intervals
Prediction Intervals
Prediction Intervals: Generate intervals to capture forecast uncertainty.
Irregular Timestamps
Irregular Timestamps
Irregular Timestamps: Process data with non-uniform timestamps directly, with no extra preprocessing.
Anomaly Detection
Anomaly Detection
Anomaly Detection: Identify anomalies automatically, integrating external features for improved precision.
Get started quickly with the Quickstart guide. Explore in-depth tutorials on TimeGPT capabilities and real-world applications.
Architecture
TimeGPT Architecture Overview
TimeGPT’s architecture builds on the self-attention mechanism introduced in the original “Attention is All You Need” paper. Unlike typical large language models (LLMs), TimeGPT is independently trained on extensive time series datasets to minimize forecasting errors.
TimeGPT employs an encoder-decoder structure with residual connections, layer normalization, and a linear output layer to match the decoder outputs to forecast dimensions. The attention-based mechanisms help the model capture diverse historical patterns to create accurate future predictions.
The model processes input sequences from left to right, similar to how humans read sentences, and predicts future values (“tokens”) based on historical windows of time series data.
Explore Examples and Use Cases
Quickstart & Setup
Quickstart & Setup
Quickly set up your workflow using our Quickstart Guide or learn to use the API by setting up your API key.
Practical Topics
Practical Topics
-
Fine-tuning with custom loss functions
-
Integrating exogenous variables, validation with cross-validation, and estimating uncertainty via quantile forecasts or prediction intervals.
Real-world Applications
Real-world Applications
With TimeGPT, you can rapidly iterate from initial exploration to high-accuracy forecasting. Dive deeper into the comprehensive tutorials for more sophisticated workflows.