metadata
dataset_info:
features:
- name: anchor
dtype: string
- name: entailment
dtype: string
- name: contradiction
dtype: string
splits:
- name: train
num_bytes: 88455406
num_examples: 551015
download_size: 39831572
dataset_size: 88455406
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
Dataset Card for "norwegian-nli-triplets"
A reformatting of compatible triplets from https://huggingface.co/datasets/tollefj/all-nli-NOB. This includes all pairs that contain both a contradiction and an entailment. In cases where a neutral also exists, this is a duplicated triplet, with the same contr./ent.
Simple normalization of sentences:
from neattext.functions import clean_text
import re
def symbol_cleaner(s):
s = re.sub(r"^[^\w\d]+", "", s)
s = re.sub(r"[^\w\d]+$", "", s)
return s
def filter_sent(sent):
sent = clean_text(sent, puncts=False, stopwords=False, non_ascii=False)
sent = symbol_cleaner(sent)
return sent
def filter_triplet(triplet):
return [filter_sent(sent) for sent in triplet]