metadata

language:
  - fr
multilinguality:
  - monolingual
task_categories:
  - token-classification
dataset_info:
  features:
    - name: tokens
      sequence: string
    - name: ner_tags_niv1
      sequence: string
    - name: ner_tags_niv2
      sequence: string
    - name: input_ids
      sequence: int32
    - name: attention_mask
      sequence: int8
    - name: labels_niv1
      sequence: int64
    - name: labels_niv2
      sequence: int64
  splits:
    - name: train
      num_bytes: 6312919
      num_examples: 6084
    - name: dev
      num_bytes: 691832
      num_examples: 676
    - name: test
      num_bytes: 1683259
      num_examples: 1685
  download_size: 1063464
  dataset_size: 8688010

m1_fine_tuning_ref_ptrn_cmbert_io

Introduction

This dataset was used to fine-tuned HueyNemud/das22-10-camembert_pretrained for nested NER task using Independant NER layers approach [M1]. It contains Paris trade directories entries from the 19th century.

Dataset parameters

Approach : M1
Dataset type : ground-truth
Tokenizer : HueyNemud/das22-10-camembert_pretrained
Tagging format : IO
Counts :
- Train : 6084
- Dev : 676
- Test : 1685
Associated fine-tuned models :
- Level-1 : nlpso/m1_ind_layers_ref_ptrn_cmbert_io_level_1
- Level 2 : nlpso/m1_ind_layers_ref_ptrn_cmbert_io_level_2

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/m1_fine_tuning_ref_ptrn_cmbert_io")