From Noisy Kenyan ID OCR to JSON with MLX and Candle

Mon, 06 Jul 2026 09:00:00 +0300

This work starts at the boundary where OCR stops. I already have a list of noisy text lines; the remaining job is to associate each value with the right field and emit one predictable JSON object:

{
 "serial_number": "...",
 "id_number": "...",
 "full_names": "...",
 "date_of_birth": "...",
 "sex": "...",
 "district_of_birth": "...",
 "place_of_issue": "...",
 "date_of_issue": "..."
}

For this experiment, I fine-tuned the model with MLX on an Apple silicon Mac, fused the adapter into the base weights, and loaded the result with Candle inside a Rust service:

Structured-Extraction on DRM HSE

From Noisy Kenyan ID OCR to JSON with MLX and Candle