[Bug]: ImportError: cannot import name 'WIKINER_FRENCH' from 'flair.datasets'

[jpmallette]

Describe the bug

Hello,

First thanks to everyone for creating this great package.

I followed https://huggingface.co/flair/ner-french tutorial to train the model and got an import error related to the dataset when running the model.
I am using flair 0.14, and python 3.11 on windows.

To Reproduce

from flair.data import Corpus
from flair.datasets import WIKINER_FRENCH
from flair.embeddings import WordEmbeddings, StackedEmbeddings, FlairEmbeddings

# 1. get the corpus
corpus: Corpus = WIKINER_FRENCH()

# 2. what tag do we want to predict?
tag_type = 'ner'

# 3. make the tag dictionary from the corpus
tag_dictionary = corpus.make_tag_dictionary(tag_type=tag_type)

# 4. initialize each embedding we use
embedding_types = [

    # GloVe embeddings
    WordEmbeddings('fr'),

    # contextual string embeddings, forward
    FlairEmbeddings('fr-forward'),

    # contextual string embeddings, backward
    FlairEmbeddings('fr-backward'),
]

# embedding stack consists of Flair and GloVe embeddings
embeddings = StackedEmbeddings(embeddings=embedding_types)

# 5. initialize sequence tagger
from flair.models import SequenceTagger

tagger = SequenceTagger(hidden_size=256,
                        embeddings=embeddings,
                        tag_dictionary=tag_dictionary,
                        tag_type=tag_type)

# 6. initialize trainer
from flair.trainers import ModelTrainer

trainer = ModelTrainer(tagger, corpus)

# 7. run training
trainer.train('resources/taggers/ner-french',
              train_with_dev=True,
              max_epochs=150)

Expected behavior

I would expect the tutorial script to work and be able to train the model.

Logs and Stack traces

from flair.datasets import WIKINER_FRENCH
ImportError: cannot import name 'WIKINER_FRENCH' from 'flair.datasets'

Screenshots

No response

Additional Context

No response

Environment

Versions:

Flair

0.14.0

Pytorch

2.5.1+cpu

Transformers

4.46.1

GPU

False

[stefan-it]

Hi @jpmallette ,

thanks for reporting that! As part of some refactoring a while ago, the new dataset class is called NER_MULTI_WIKINER and the French part can be loaded with:

from flair.datasets import NER_MULTI_WIKINER

corpus = NER_MULTI_WIKINER(languages="fr")
print(corpus)

This will output:

MultiCorpus: 107128 train + 11903 dev + 13226 test sentences
 - ColumnCorpus Corpus: 107128 train + 11903 dev + 13226 test sentences - /home/stefan/.flair/datasets/ner_multi_wikiner/fr

These numbers do match with the training log of that ner-french model: https://huggingface.co/flair/ner-french/blob/main/training.log#L32-L33

I submitted a PR for the model card here, cc @alanakbik for review :)

[alanakbik]

Thanks a lot @stefan-it! Just merged your PR!

[jpmallette]

Awesome, thanks to @stefan-it @alanakbik for your prompt reply!