metadata

language:
  - fr
multilinguality:
  - monolingual
task_categories:
  - token-classification

m2m3_qualitative_analysis_ocr_cmbert_iob2

Introduction

This dataset was used to perform qualitative analysis of Jean-Baptiste/camembert-ner on nested NER task using Independant NER layers approach [M1]. It contains Paris trade directories entries from the 19th century.

Dataset parameters

Approachrd : M2 and M3
Dataset type : noisy (Pero OCR)
Tokenizer : Jean-Baptiste/camembert-ner
Tagging format : IOB2
Counts :
- Train : 6084
- Dev : 676
- Test : 1685
Associated fine-tuned models :
- M2 : nlpso/m2_joint_label_ocr_cmbert_iob2
- M3 : nlpso/m3_hierarchical_ner_ocr_cmbert_iob2

Entity types

Abbreviation	Entity group (level)	Description
O	1 & 2	Outside of a named entity
PER	1	Person or company name
ACT	1 & 2	Person or company professional activity
TITREH	2	Military or civil distinction
DESC	1	Entry full description
TITREP	2	Professionnal reward
SPAT	1	Address
LOC	2	Street name
CARDINAL	2	Street number
FT	2	Geographical feature

How to use this dataset

from datasets import load_dataset

train_dev_test = load_dataset("nlpso/m2m3_qualitative_analysis_ocr_cmbert_iob2")