Back to walidsassi.com
Scoring a multimodal receipt extraction with code metrics and a model judge

Scoring What the Model Read: Metrics, a Judge, and 21 Receipts

Comparing store names with == fails almost every receipt. Here are the three code metrics and the model judge that turn a photo extraction into numbers, plus the results on 21 real receipts and the error patterns hiding behind them.

September 12, 2026 · 10 min · Walid Sassi