You will learn the evaluation steps — collecting development data, defining DSPy metrics, and running evaluations — while become proficient in data handling with Example objects and creating metrics to assess output quality.
Overview
Syllabus
- Unit 1: Steps in Evaluation
- DSPy Evaluation Steps Quiz
- Unit 2: Data Handling with Example Objects in DSPy
- Creating and Exploring Example Objects
- Building a Mini Dataset with Examples
- Separating Inputs and Labels Programmatically
- Filtering Example Data by Specific Criteria
- Transforming Example Data While Preserving Structure
- Unit 3: Creating Metrics in DSPy
- Implementing a Case Insensitive Metric
- Flexible Answer Matching for Multiple Formats
- Passage Matching for Retrieval Systems
- Building a Holistic QA Evaluation Metric
- Semantic Evaluation with F1 Score