Accuracy Eval Cookbooks

Accuracy examples evaluate how well responses match expected outputs.

Files

• accuracy_basic.py - Sync and async calculator accuracy evaluations.
• accuracy_9_11_bigger_or_9_99.py - Numeric comparison accuracy evaluation.
• accuracy_team.py - Team language-routing accuracy evaluation.
• accuracy_with_given_answer.py - Accuracy scoring for a provided output string.
• accuracy_with_tools.py - Accuracy evaluation for a tool-using agent.
• db_logging.py - Accuracy evaluation with PostgreSQL result logging.
• evaluator_agent.py - Accuracy evaluation using a custom evaluator agent.
• accuracy_eval_metrics.py - Eval model metrics accumulated into agent run_output under "eval_model" detail key.