TOTO<T>

Models & Types · AiDotNet.Finance.Forecasting.Foundation

TOTO — Datadog’s Time Series Foundation Model for Observability.

Inheritance: NeuralNetworkBase<T> → FinancialModelBase<T> → ForecastingModelBase<T> → TimeSeriesFoundationModelBase<T> → TOTO<T>

For Beginners

TOTO is a specialized model from Datadog, trained on 1 trillion data points from real-world server monitoring. It excels at predicting system metrics like CPU usage, memory consumption, and request latency. If you monitor infrastructure and need to forecast capacity or detect anomalies before they cause outages, TOTO is purpose-built for that domain.

How It Works

TOTO is Datadog’s domain-specific time series foundation model optimized for IT operations, infrastructure monitoring, and observability. Pre-trained on 1 trillion data points from the Datadog observability platform, it excels at SRE metrics and anomaly detection.

Reference: Datadog, “Introducing Toto: A state-of-the-art time series foundation model”, 2025.

Example

// Create a TOTO model (Datadog) for observability time series forecasting
// Pre-trained on 1 trillion data points from real-world infrastructure monitoring
var architecture = new NeuralNetworkArchitecture<double>(
    inputType: InputType.OneDimensional,
    taskType: NeuralNetworkTaskType.Regression,
    inputHeight: 512, inputWidth: 1, inputDepth: 1, outputSize: 24);
            
// Training mode optimized for SRE metrics and anomaly detection
var model = new TOTO<double>(architecture);
            
// ONNX inference mode with pre-trained model
var onnxModel = new TOTO<double>(architecture, "toto.onnx");

Constructors

Constructor Summary
TOTO(NeuralNetworkArchitecture<T> architecture, TOTOOptions<T> options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer, ILossFunction<T> lossFunction) Creates a TOTO model in native mode for training or fine-tuning.
TOTO(NeuralNetworkArchitecture<T> architecture, string onnxModelPath, TOTOOptions<T> options, IGradientBasedOptimizer<T, Tensor<T>, Tensor<T>> optimizer, ILossFunction<T> lossFunction) Creates a TOTO model using a pretrained ONNX model.

Properties

Property Summary
FoundationModelSize ModelSize { get; }
bool IsChannelIndependent { get; }
bool SupportsTraining { get; }
bool UseNativeMode { get; }
int MaxContextLength { get; }
int MaxPredictionHorizon { get; }
int NumFeatures { get; }
int PatchSize { get; }
int PredictionHorizon { get; }
int SequenceLength { get; }
int Stride { get; }

Methods

Method Summary
Dictionary<string, T> Evaluate(Tensor<T> predictions, Tensor<T> actuals)
Dictionary<string, T> GetFinancialMetrics()
Dictionary<string, Tensor<T>> GetNamedLayerActivations(Tensor<T> input)
ModelMetadata<T> GetModelMetadata()
Tensor<T> ApplyInstanceNormalization(Tensor<T> input)
Tensor<T> AutoregressiveForecast(Tensor<T> input, int steps)