Interactive visualization of Datadog's Time-Series-Optimized Transformer for Observability metrics. Explore zero-shot forecasting, probabilistic predictions, and multi-variate time series analysis.
TOTO can forecast time series without fine-tuning on your specific data. Click on the chart to add anomalies!
TOTO efficiently processes multiple variables using Proportional Factorized Space-Time Attention. Hover over lines to see correlations!
Experiment with TOTO's architecture parameters and see how they affect performance.
Pre-trained on 2+ trillion time series data points, the largest dataset for any open-weights time series model.
Specifically designed for monitoring metrics covering infrastructure, networking, databases, and applications.
Generates both point forecasts and uncertainty estimates using a Student-T mixture model.
State-of-the-art results on GIFT-Eval benchmark and custom BOOM observability dataset.
Supports variable prediction horizons and context lengths for flexible forecasting.
Available on Hugging Face with optimizations for xformers and flash-attention.