name: sequence_labeling
config_type: dataset
task: sequence_labeling
path: hezarai/arman-ner
tokenizer_path: hezarai/distilbert-base-fa
tags_field: ner_tags
tokens_field: tokens
ignore_index: -100
label_all_tokens: true