ModelPredictionTool

Models & Types · AiDotNet.Agentic.Tools

Exposes a trained model’s prediction as a callable agent tool, so an LLM agent can query the model you just trained mid-reasoning — the same way it would call web search or a calculator.

Inheritance: AgentToolBase → ModelPredictionTool

For Beginners

This turns your trained model into a “gadget” the assistant can use. When the assistant needs a prediction, it calls this tool with the input numbers and gets the model’s answer back.

How It Works

The tool takes a numeric feature vector as { "features": [ ... ] } and returns the model’s output vector as JSON. It is built by AiModelResult for tabular models (matrix/vector shaped), which supplies the marshalling delegate; unsupported model shapes simply do not produce this tool.

Constructors

Constructor Summary
ModelPredictionTool(Func<double[], double[]> predict, int featureCount, string name, string description) Creates a model-prediction tool.

Methods

Method Summary
double? TryParseFirstPrediction(string toolResultContent) Parses the first numeric value from this tool’s JSON result, for grounded verification.