French NER model for tweets. Fine-tuned on the CAP2017 dataset. 

label_list = ['O',
                 'B-person',
                 'I-person',
                 'B-musicartist',
                 'I-musicartist',
                 'B-org',
                 'I-org',
                 'B-geoloc',
                 'I-geoloc',
                 'B-product',
                 'I-product',
                 'B-transportLine',
                 'I-transportLine',
                 'B-media',
                 'I-media',
                 'B-sportsteam',
                 'I-sportsteam',
                 'B-event',
                 'I-event',
                 'B-tvshow',
                 'I-tvshow',
                 'B-movie',
                 'I-movie',
                 'B-facility',
                 'I-facility',
                 'B-other',
                 'I-other']