Add the pretrained weights

8cbed27 verified 2 days ago

1.8 kB

tags:
  - model_hub_mixin
  - pytorch_model_hub_mixin
  - address-parsing
  - deepparse
license: mit
datasets:
  - deepparse/worldwide-addresses
library_name: deepparse

deepparse model2vec-8m

A deepparse address parser (Seq2Seq, without attention mechanism) using the minishlab/potion-base-8M Model2Vec static embeddings (256 dimensions).

from deepparse.parser import AddressParser

address_parser = AddressParser(model_type="model2vec-8m")
address_parser("350 rue des Lilas Ouest Québec Québec G1L 1B6")

Install it with pip install deepparse[model2vec].

Training

Data: the deepparse/worldwide-addresses dataset (20 countries, 100,000 clean and 50,000 incomplete addresses per country), 90% for training and 10% for validation. Suburb and District are mapped to Municipality, and the County and Country words are removed.
Architecture: encoder and decoder hidden size 1024, one layer each.
Selection: the hidden size (128, 256, 512 or 1024) with the lowest mean validation loss over 5 seeds, then the seed with the lowest validation loss (seed 42, best epoch 29).

Results

	Accuracy (%)
Test, this model	97.95
Test, mean ± std over the 5 seeds of this configuration	97.78 ± 0.32
Validation, this model	98.21

The accuracy is the per-word tag accuracy on the test split of the rebuilt worldwide-addresses data.

Limitations

The source data has no Orientation nor GeneralDelivery tags: the model has these outputs but never predicts them (the street orientation, e.g. Ouest or NW, is part of the StreetName). South Korea has almost only incomplete addresses.