Skip to main content

v0.1.0 PyTorch checkpoint × golden v0.1.2 (apples-to-apples baseline)

  • entries evaluated: 4535
  • full-parseaddress parsingThe process of decomposing a free-text postal address string into structured components — house number, street name, locality, region, postcode, and country — so a geocoder can resolve them to coordinates. exact matchexact matchThe share of eval items whose every component is correct (compared per-span or per-token). Stricter than per-tag F1, which credits partial correctness.: 0.0088
  • mean tokentokenOne word or subword in the tokenized input. For the neural classifier, tokens come from SentencePiece (subword units); for the rule classifiers, tokens are whitespace- and punctuation-separated words. confidence: 0.8587

Per-component F1

tagprecisionprecisionOf the spans the model labeled as a given tag, the fraction it got right. High precision means few false positives. Paired with recall to compute F1.recallrecallOf the spans whose gold label is a given tag, the fraction the model found. High recall means few misses. Paired with precision to compute F1.f1support
countrycountryThe top-level address component (an ISO country). Closed-vocabulary, so it is best handled by a deterministic matcher feeding a proposal rather than a retrained model head.0.01350.13880.0246245
regionregionThe first-level administrative subdivision of a country — a US state, a French region, a province. The component between country and locality.0.03330.04090.03673205
localitylocalityThe city / town / settlement component of an address: a populated place sitting between region and neighbourhood in the hierarchy.0.06770.09090.07763357
dependent_locality0.00000.00000.000040
postcodepostcodeThe country-specific postal code (US ZIP, French code postal, etc.). Mailwoman handles postcode parsing entirely by rule classifier — a regex problem, not an ML one.1.00000.00030.00072980
subregion0.00000.00000.00000
cedexCEDEX (Courrier d'Entreprise à Distribution Exceptionnelle). A French postal routing for high-volume business mail: a CEDEX code delivers directly from a sorting centre, bypassing the local post office. A common negative-space format Mailwoman must parse.0.00000.00000.00001

Calibration (confidence bucket → accuracy)

bucketnaccuracy
0.0–0.100.0000
0.1–0.211.0000
0.2–0.32080.0913
0.3–0.416600.1084
0.4–0.532070.1366
0.5–0.640300.1846
0.6–0.739810.2042
0.7–0.842060.2183
0.8–0.956270.2472
0.9–1.0383280.3366