IGBO_NER addition to FlairNLP

[PhilippThamm]

Added the corpus IGBO_NER for named entity recognition in the language Igbo to the flair framework.

[alanakbik]

@PhilippThamm I am getting some weird results when I load this dataset:

corpus = IGBO_NER()

# print statistics
print(corpus)

# print some example sentences
print(corpus.train[0])
print(corpus.train[0])
print(corpus.train[1])
print(corpus.train[2])
print(corpus.train[3])

# print the NER dictionary (should be only NER tags like B-PER, I-PER, E-PER, B-ORG, I-ORG)
corpus.make_label_dictionary('ner')

This prints a lot of weird characters. Perhaps the encoding is wrongly set?

[alanakbik]

Fix encoding and set tag_to_bioes variable

[Dabendorf]

@PhilippThamm I am getting some weird results when I load this dataset:

corpus = IGBO_NER()

# print statistics
print(corpus)

# print some example sentences
print(corpus.train[0])
print(corpus.train[0])
print(corpus.train[1])
print(corpus.train[2])
print(corpus.train[3])

# print the NER dictionary (should be only NER tags like B-PER, I-PER, E-PER, B-ORG, I-ORG)
corpus.make_label_dictionary('ner')

This prints a lot of weird characters. Perhaps the encoding is wrongly set?

As I understand it, the encoding should be set to utf-8 instead of Latin-1

[alanakbik]

@PhilippThamm thanks for adding this!