dspy-evaluation-suite

Evaluate DSPy programs with custom metrics and parallel execution.

120|13|Updated Dec 21, 2025
One-click install
npx skills add https://github.com/OmidZamani/dspy-skills --skill dspy-evaluation-suite
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: dspy-evaluation-suite
Source: https://github.com/OmidZamani/dspy-skills/tree/main/skills/dspy-evaluation-suite
Command: npx skills add https://github.com/OmidZamani/dspy-skills --skill dspy-evaluation-suite

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides a framework to measure DSPy program performance using built-in and custom metrics with parallel execution.

Core Features & Use Cases

  • Built-in metrics such as answer_exact_match and SemanticF1
  • Custom metrics with multi-factor scoring
  • Production-grade evaluation workflow and reporting

Quick Start

Instantiate Evaluate with a devset and metric, then run evaluator(program) to obtain a score and results; optionally compare multiple programs and export reports.

Frequently Asked Questions about dspy-evaluation-suite

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I evaluate DSPy programs to measure their performance?

Evaluate DSPy programs by instantiating the Evaluate component with your development set and a scoring function, then run the evaluator on your program to obtain a numeric score and per-example results. Built-in metrics like answer_exact_match and SemanticF1 are available, or supply custom metrics for multi-factor scoring.

What metrics can I use to benchmark DSPy program variants?

Use built-in metrics including answer_exact_match and SemanticF1, or define custom metrics with multi-factor scoring logic. The evaluation suite applies your chosen metric across all examples in parallel, producing comparable scores for variant comparison and production-readiness validation.

Can I run parallel evaluations across multiple DSPy programs?

Yes. The evaluation suite supports configurable parallel execution with threaded evaluation across your development set. This enables efficient benchmarking and baseline establishment when comparing multiple program variants simultaneously.

What input format does the DSPy evaluation suite require?

Provide a development set and a pluggable scoring function. The suite accepts both built-in metrics and custom scoring functions, then produces structured output containing a numeric aggregate score and detailed per-example results for analysis.

When should I use this evaluation framework instead of manual testing?

Use this framework when you need reproducible, quantified performance measurement across DSPy program variants, especially for baseline establishment, production-readiness validation, and scaled evaluation where manual inspection becomes impractical.

How do I export evaluation results and reports?

After running evaluation, the suite returns structured results with numeric scores and per-example breakdowns that can be exported for reporting. Compare multiple program runs to track performance across iterations and document evaluation findings.