|
"""madlad-400 dataset""" |
|
|
|
|
|
import gzip |
|
import json |
|
|
|
import datasets |
|
|
|
|
|
logger = datasets.logging.get_logger(__name__) |
|
|
|
|
|
_DESCRIPTION = """\ |
|
MADLAD-400 (Multilingual Audited Dataset: Low-resource And Document-level) is a document-level multilingual dataset based on Common Crawl, covering 419 languages in total. We use all snapshots of CommonCrawl available as of August 1, 2022. The primary advantage of this dataset over similar datasets is that it is more multilingual (419 languages), it is audited and more highly filtered, and it is document-level. The main disadvantage is also its strength -- being more filtered, it may lack the recall needed for some applications. |
|
""" |
|
|
|
_CITATION = """ |
|
@misc{kudugunta2023madlad400, |
|
title={MADLAD-400: A Multilingual And Document-Level Large Audited Dataset}, |
|
author={Sneha Kudugunta and Isaac Caswell and Biao Zhang and Xavier Garcia and Christopher A. Choquette-Choo and Katherine Lee and Derrick Xin and Aditya Kusupati and Romi Stella and Ankur Bapna and Orhan Firat}, |
|
year={2023}, |
|
eprint={2309.04662}, |
|
archivePrefix={arXiv}, |
|
primaryClass={cs.CL} |
|
} |
|
""" |
|
|
|
_DATA_URL = "https://huggingface.co/datasets/allenai/MADLAD-400/resolve/ecd71297d60c1eb996cd3d7c44c60ad5b55adfc6/data/{language}/{language}_{split}_{index:04d}.jsonl.gz" |
|
|
|
|
|
|
|
|
|
_N_SHARDS_PER_SPLIT = { |
|
'abt': {'clean': 1, 'noisy': 1}, |
|
'ace': {'clean': 1, 'noisy': 1}, |
|
'acf': {'clean': 1, 'noisy': 1}, |
|
'ach': {'clean': 1, 'noisy': 1}, |
|
'ada': {'clean': 1, 'noisy': 1}, |
|
'adh': {'clean': 1, 'noisy': 1}, |
|
'ady': {'clean': 1, 'noisy': 1}, |
|
'af': {'clean': 1, 'noisy': 2}, |
|
'agr': {'clean': 1, 'noisy': 1}, |
|
'ahk': {'clean': 1, 'noisy': 1}, |
|
'ak': {'clean': 1, 'noisy': 1}, |
|
'akb': {'clean': 1, 'noisy': 1}, |
|
'alt': {'clean': 1, 'noisy': 1}, |
|
'alz': {'clean': 1, 'noisy': 1}, |
|
'am': {'clean': 1, 'noisy': 1}, |
|
'amu': {'clean': 1, 'noisy': 1}, |
|
'ang': {'clean': 1, 'noisy': 1}, |
|
'ann': {'clean': 1, 'noisy': 1}, |
|
'ape': {'clean': 1, 'noisy': 1}, |
|
'ar': {'clean': 8, 'noisy': 44}, |
|
'arn': {'clean': 1, 'noisy': 1}, |
|
'as': {'clean': 1, 'noisy': 1}, |
|
'av': {'clean': 1, 'noisy': 1}, |
|
'awa': {'clean': 1, 'noisy': 1}, |
|
'ay': {'clean': 1, 'noisy': 1}, |
|
'az': {'clean': 2, 'noisy': 3}, |
|
'az_RU': {'clean': 1, 'noisy': 1}, |
|
'ba': {'clean': 1, 'noisy': 1}, |
|
'ban': {'clean': 1, 'noisy': 1}, |
|
'bas': {'clean': 1, 'noisy': 1}, |
|
'bbc': {'clean': 1, 'noisy': 1}, |
|
'bci': {'clean': 1, 'noisy': 1}, |
|
'be': {'clean': 1, 'noisy': 2}, |
|
'ber': {'clean': 1, 'noisy': 1}, |
|
'ber_Latn': {'clean': 1, 'noisy': 1}, |
|
'bew': {'clean': 1, 'noisy': 1}, |
|
'bg': {'clean': 10, 'noisy': 16}, |
|
'bg_Latn': {'clean': 1, 'noisy': 1}, |
|
'bgp': {'clean': 1, 'noisy': 1}, |
|
'bho': {'clean': 1, 'noisy': 1}, |
|
'bi': {'clean': 1, 'noisy': 1}, |
|
'bik': {'clean': 1, 'noisy': 1}, |
|
'bim': {'clean': 1, 'noisy': 1}, |
|
'bm': {'clean': 1, 'noisy': 1}, |
|
'bn': {'clean': 1, 'noisy': 4}, |
|
'bn_Latn': {'clean': 1, 'noisy': 1}, |
|
'bo': {'clean': 1, 'noisy': 1}, |
|
'bqc': {'clean': 1, 'noisy': 1}, |
|
'br': {'clean': 1, 'noisy': 1}, |
|
'bru': {'clean': 1, 'noisy': 1}, |
|
'brx': {'clean': 1, 'noisy': 1}, |
|
'bs': {'clean': 1, 'noisy': 6}, |
|
'bts': {'clean': 1, 'noisy': 1}, |
|
'btx': {'clean': 1, 'noisy': 1}, |
|
'bua': {'clean': 1, 'noisy': 1}, |
|
'bum': {'clean': 1, 'noisy': 1}, |
|
'bus': {'clean': 1, 'noisy': 1}, |
|
'bzj': {'clean': 1, 'noisy': 1}, |
|
'ca': {'clean': 5, 'noisy': 8}, |
|
'cab': {'clean': 1, 'noisy': 1}, |
|
'cac': {'clean': 1, 'noisy': 1}, |
|
'cak': {'clean': 1, 'noisy': 1}, |
|
'cbk': {'clean': 1, 'noisy': 1}, |
|
'cce': {'clean': 1, 'noisy': 1}, |
|
'ce': {'clean': 1, 'noisy': 1}, |
|
'ceb': {'clean': 1, 'noisy': 1}, |
|
'cfm': {'clean': 1, 'noisy': 1}, |
|
'ch': {'clean': 1, 'noisy': 1}, |
|
'chk': {'clean': 1, 'noisy': 1}, |
|
'chm': {'clean': 1, 'noisy': 1}, |
|
'chr': {'clean': 1, 'noisy': 1}, |
|
'ckb': {'clean': 1, 'noisy': 1}, |
|
'cnh': {'clean': 1, 'noisy': 1}, |
|
'co': {'clean': 1, 'noisy': 1}, |
|
'cr_Latn': {'clean': 1, 'noisy': 1}, |
|
'crh': {'clean': 1, 'noisy': 1}, |
|
'crs': {'clean': 1, 'noisy': 1}, |
|
'cs': {'clean': 23, 'noisy': 44}, |
|
'ctd_Latn': {'clean': 1, 'noisy': 1}, |
|
'ctu': {'clean': 1, 'noisy': 1}, |
|
'cuk': {'clean': 1, 'noisy': 1}, |
|
'cv': {'clean': 1, 'noisy': 1}, |
|
'cy': {'clean': 1, 'noisy': 4}, |
|
'da': {'clean': 11, 'noisy': 36}, |
|
'de': {'clean': 133, 'noisy': 293}, |
|
'din': {'clean': 1, 'noisy': 1}, |
|
'dje': {'clean': 1, 'noisy': 1}, |
|
'djk': {'clean': 1, 'noisy': 1}, |
|
'dln': {'clean': 1, 'noisy': 1}, |
|
'doi': {'clean': 1, 'noisy': 1}, |
|
'dov': {'clean': 1, 'noisy': 1}, |
|
'dtp': {'clean': 1, 'noisy': 1}, |
|
'dv': {'clean': 1, 'noisy': 1}, |
|
'dwr': {'clean': 1, 'noisy': 1}, |
|
'dyu': {'clean': 1, 'noisy': 1}, |
|
'dz': {'clean': 1, 'noisy': 1}, |
|
'ee': {'clean': 1, 'noisy': 1}, |
|
'el': {'clean': 14, 'noisy': 29}, |
|
'el_Latn': {'clean': 1, 'noisy': 1}, |
|
'emp': {'clean': 1, 'noisy': 1}, |
|
'en': {'clean': 947, 'noisy': 1555}, |
|
'enq': {'clean': 1, 'noisy': 1}, |
|
'eo': {'clean': 1, 'noisy': 1}, |
|
'es': {'clean': 137, 'noisy': 269}, |
|
'et': {'clean': 4, 'noisy': 6}, |
|
'eu': {'clean': 1, 'noisy': 1}, |
|
'fa': {'clean': 16, 'noisy': 35}, |
|
'ff': {'clean': 1, 'noisy': 1}, |
|
'ffm': {'clean': 1, 'noisy': 1}, |
|
'fi': {'clean': 14, 'noisy': 27}, |
|
'fil': {'clean': 1, 'noisy': 2}, |
|
'fip': {'clean': 1, 'noisy': 1}, |
|
'fj': {'clean': 1, 'noisy': 1}, |
|
'fo': {'clean': 1, 'noisy': 1}, |
|
'fon': {'clean': 1, 'noisy': 1}, |
|
'fr': {'clean': 133, 'noisy': 251}, |
|
'frp': {'clean': 1, 'noisy': 1}, |
|
'fy': {'clean': 1, 'noisy': 1}, |
|
'ga': {'clean': 1, 'noisy': 4}, |
|
'gag': {'clean': 1, 'noisy': 1}, |
|
'gbm': {'clean': 1, 'noisy': 1}, |
|
'gd': {'clean': 1, 'noisy': 1}, |
|
'gl': {'clean': 1, 'noisy': 2}, |
|
'gn': {'clean': 1, 'noisy': 1}, |
|
'gof': {'clean': 1, 'noisy': 1}, |
|
'gom': {'clean': 1, 'noisy': 1}, |
|
'gom_Latn': {'clean': 1, 'noisy': 1}, |
|
'gor': {'clean': 1, 'noisy': 1}, |
|
'grc': {'clean': 1, 'noisy': 1}, |
|
'gsw': {'clean': 1, 'noisy': 5}, |
|
'gu': {'clean': 1, 'noisy': 1}, |
|
'gu_Latn': {'clean': 1, 'noisy': 1}, |
|
'gub': {'clean': 1, 'noisy': 1}, |
|
'guc': {'clean': 1, 'noisy': 1}, |
|
'guh': {'clean': 1, 'noisy': 1}, |
|
'gui': {'clean': 1, 'noisy': 1}, |
|
'gv': {'clean': 1, 'noisy': 1}, |
|
'gvl': {'clean': 1, 'noisy': 1}, |
|
'gym': {'clean': 1, 'noisy': 1}, |
|
'ha': {'clean': 1, 'noisy': 1}, |
|
'haw': {'clean': 1, 'noisy': 1}, |
|
'he': {'clean': 6, 'noisy': 10}, |
|
'hi': {'clean': 4, 'noisy': 8}, |
|
'hi_Latn': {'clean': 1, 'noisy': 1}, |
|
'hif': {'clean': 1, 'noisy': 2}, |
|
'hil': {'clean': 1, 'noisy': 1}, |
|
'hmn': {'clean': 1, 'noisy': 1}, |
|
'hne': {'clean': 1, 'noisy': 1}, |
|
'ho': {'clean': 1, 'noisy': 1}, |
|
'hr': {'clean': 2, 'noisy': 13}, |
|
'ht': {'clean': 1, 'noisy': 1}, |
|
'hu': {'clean': 20, 'noisy': 34}, |
|
'hui': {'clean': 1, 'noisy': 1}, |
|
'hus': {'clean': 1, 'noisy': 1}, |
|
'hvn': {'clean': 1, 'noisy': 1}, |
|
'hy': {'clean': 1, 'noisy': 2}, |
|
'iba': {'clean': 1, 'noisy': 1}, |
|
'ibb': {'clean': 1, 'noisy': 1}, |
|
'id': {'clean': 18, 'noisy': 54}, |
|
'ify': {'clean': 1, 'noisy': 1}, |
|
'ig': {'clean': 1, 'noisy': 1}, |
|
'ilo': {'clean': 1, 'noisy': 1}, |
|
'inb': {'clean': 1, 'noisy': 1}, |
|
'is': {'clean': 1, 'noisy': 3}, |
|
'iso': {'clean': 1, 'noisy': 1}, |
|
'it': {'clean': 72, 'noisy': 170}, |
|
'iu': {'clean': 1, 'noisy': 1}, |
|
'ium': {'clean': 1, 'noisy': 1}, |
|
'izz': {'clean': 1, 'noisy': 1}, |
|
'ja': {'clean': 36, 'noisy': 36}, |
|
'jac': {'clean': 1, 'noisy': 1}, |
|
'jam': {'clean': 1, 'noisy': 1}, |
|
'jiv': {'clean': 1, 'noisy': 1}, |
|
'jv': {'clean': 1, 'noisy': 1}, |
|
'jvn': {'clean': 1, 'noisy': 1}, |
|
'ka': {'clean': 1, 'noisy': 2}, |
|
'kaa': {'clean': 1, 'noisy': 1}, |
|
'kaa_Latn': {'clean': 1, 'noisy': 1}, |
|
'kac': {'clean': 1, 'noisy': 1}, |
|
'kbd': {'clean': 1, 'noisy': 1}, |
|
'kbp': {'clean': 1, 'noisy': 1}, |
|
'kek': {'clean': 1, 'noisy': 1}, |
|
'kg': {'clean': 1, 'noisy': 1}, |
|
'kha': {'clean': 1, 'noisy': 1}, |
|
'kj': {'clean': 1, 'noisy': 1}, |
|
'kjg': {'clean': 1, 'noisy': 1}, |
|
'kjh': {'clean': 1, 'noisy': 1}, |
|
'kk': {'clean': 2, 'noisy': 3}, |
|
'kl': {'clean': 1, 'noisy': 1}, |
|
'km': {'clean': 1, 'noisy': 1}, |
|
'kmb': {'clean': 1, 'noisy': 1}, |
|
'kmz_Latn': {'clean': 1, 'noisy': 1}, |
|
'kn': {'clean': 1, 'noisy': 1}, |
|
'kn_Latn': {'clean': 1, 'noisy': 1}, |
|
'knj': {'clean': 1, 'noisy': 1}, |
|
'ko': {'clean': 13, 'noisy': 19}, |
|
'koi': {'clean': 1, 'noisy': 1}, |
|
'kos': {'clean': 1, 'noisy': 1}, |
|
'krc': {'clean': 1, 'noisy': 1}, |
|
'kri': {'clean': 1, 'noisy': 1}, |
|
'ksd': {'clean': 1, 'noisy': 1}, |
|
'ksw': {'clean': 1, 'noisy': 1}, |
|
'ktu': {'clean': 1, 'noisy': 1}, |
|
'ku': {'clean': 1, 'noisy': 1}, |
|
'kum': {'clean': 1, 'noisy': 1}, |
|
'kv': {'clean': 1, 'noisy': 1}, |
|
'kw': {'clean': 1, 'noisy': 1}, |
|
'kwi': {'clean': 1, 'noisy': 1}, |
|
'ky': {'clean': 1, 'noisy': 1}, |
|
'la': {'clean': 1, 'noisy': 2}, |
|
'laj': {'clean': 1, 'noisy': 1}, |
|
'lb': {'clean': 1, 'noisy': 8}, |
|
'lg': {'clean': 1, 'noisy': 1}, |
|
'lhu': {'clean': 1, 'noisy': 1}, |
|
'ln': {'clean': 1, 'noisy': 1}, |
|
'lo': {'clean': 1, 'noisy': 1}, |
|
'lrc': {'clean': 1, 'noisy': 1}, |
|
'lt': {'clean': 6, 'noisy': 9}, |
|
'ltg': {'clean': 1, 'noisy': 1}, |
|
'lu': {'clean': 1, 'noisy': 1}, |
|
'lus': {'clean': 1, 'noisy': 1}, |
|
'lv': {'clean': 4, 'noisy': 5}, |
|
'mad': {'clean': 1, 'noisy': 1}, |
|
'mag': {'clean': 1, 'noisy': 1}, |
|
'mai': {'clean': 1, 'noisy': 1}, |
|
'mak': {'clean': 1, 'noisy': 1}, |
|
'mam': {'clean': 1, 'noisy': 1}, |
|
'mas': {'clean': 1, 'noisy': 1}, |
|
'maz': {'clean': 1, 'noisy': 1}, |
|
'mbt': {'clean': 1, 'noisy': 1}, |
|
'mdf': {'clean': 1, 'noisy': 1}, |
|
'meo': {'clean': 1, 'noisy': 1}, |
|
'meu': {'clean': 1, 'noisy': 1}, |
|
'mfe': {'clean': 1, 'noisy': 1}, |
|
'mg': {'clean': 1, 'noisy': 1}, |
|
'mgh': {'clean': 1, 'noisy': 1}, |
|
'mh': {'clean': 1, 'noisy': 1}, |
|
'mi': {'clean': 1, 'noisy': 1}, |
|
'min': {'clean': 1, 'noisy': 1}, |
|
'miq': {'clean': 1, 'noisy': 1}, |
|
'mk': {'clean': 1, 'noisy': 2}, |
|
'mkn': {'clean': 1, 'noisy': 1}, |
|
'ml': {'clean': 2, 'noisy': 2}, |
|
'ml_Latn': {'clean': 1, 'noisy': 1}, |
|
'mn': {'clean': 1, 'noisy': 2}, |
|
'mni': {'clean': 1, 'noisy': 1}, |
|
'mps': {'clean': 1, 'noisy': 1}, |
|
'mqy': {'clean': 1, 'noisy': 1}, |
|
'mr': {'clean': 1, 'noisy': 2}, |
|
'mrj': {'clean': 1, 'noisy': 1}, |
|
'mrw': {'clean': 1, 'noisy': 1}, |
|
'ms': {'clean': 2, 'noisy': 8}, |
|
'ms_Arab': {'clean': 1, 'noisy': 1}, |
|
'ms_Arab_BN': {'clean': 1, 'noisy': 1}, |
|
'msb': {'clean': 1, 'noisy': 1}, |
|
'msi': {'clean': 1, 'noisy': 1}, |
|
'msm': {'clean': 1, 'noisy': 1}, |
|
'mt': {'clean': 1, 'noisy': 1}, |
|
'my': {'clean': 1, 'noisy': 1}, |
|
'myv': {'clean': 1, 'noisy': 1}, |
|
'nan_Latn_TW': {'clean': 1, 'noisy': 1}, |
|
'ndc_ZW': {'clean': 1, 'noisy': 1}, |
|
'ne': {'clean': 1, 'noisy': 1}, |
|
'new': {'clean': 1, 'noisy': 1}, |
|
'ngu': {'clean': 1, 'noisy': 1}, |
|
'nhe': {'clean': 1, 'noisy': 1}, |
|
'nia': {'clean': 1, 'noisy': 1}, |
|
'nij': {'clean': 1, 'noisy': 1}, |
|
'niq': {'clean': 1, 'noisy': 1}, |
|
'nl': {'clean': 43, 'noisy': 95}, |
|
'nnb': {'clean': 1, 'noisy': 1}, |
|
'no': {'clean': 10, 'noisy': 40}, |
|
'noa': {'clean': 1, 'noisy': 1}, |
|
'nog': {'clean': 1, 'noisy': 1}, |
|
'nr': {'clean': 1, 'noisy': 1}, |
|
'nso': {'clean': 1, 'noisy': 1}, |
|
'nut': {'clean': 1, 'noisy': 1}, |
|
'nv': {'clean': 1, 'noisy': 1}, |
|
'ny': {'clean': 1, 'noisy': 1}, |
|
'nyu': {'clean': 1, 'noisy': 1}, |
|
'nzi': {'clean': 1, 'noisy': 1}, |
|
'oc': {'clean': 1, 'noisy': 1}, |
|
'oj': {'clean': 1, 'noisy': 1}, |
|
'om': {'clean': 1, 'noisy': 1}, |
|
'or': {'clean': 1, 'noisy': 1}, |
|
'os': {'clean': 1, 'noisy': 1}, |
|
'otq': {'clean': 1, 'noisy': 1}, |
|
'pa': {'clean': 1, 'noisy': 1}, |
|
'pag': {'clean': 1, 'noisy': 1}, |
|
'pap': {'clean': 1, 'noisy': 1}, |
|
'pau': {'clean': 1, 'noisy': 1}, |
|
'pck': {'clean': 1, 'noisy': 1}, |
|
'pis': {'clean': 1, 'noisy': 1}, |
|
'pl': {'clean': 51, 'noisy': 74}, |
|
'pon': {'clean': 1, 'noisy': 1}, |
|
'ppk': {'clean': 1, 'noisy': 1}, |
|
'ps': {'clean': 1, 'noisy': 1}, |
|
'pt': {'clean': 66, 'noisy': 104}, |
|
'qu': {'clean': 1, 'noisy': 1}, |
|
'qub': {'clean': 1, 'noisy': 1}, |
|
'quc': {'clean': 1, 'noisy': 1}, |
|
'quf': {'clean': 1, 'noisy': 1}, |
|
'quh': {'clean': 1, 'noisy': 1}, |
|
'qup': {'clean': 1, 'noisy': 1}, |
|
'quy': {'clean': 1, 'noisy': 1}, |
|
'qvc': {'clean': 1, 'noisy': 1}, |
|
'qvi': {'clean': 1, 'noisy': 1}, |
|
'qvz': {'clean': 1, 'noisy': 1}, |
|
'qxr': {'clean': 1, 'noisy': 1}, |
|
'raj': {'clean': 1, 'noisy': 1}, |
|
'rcf': {'clean': 1, 'noisy': 1}, |
|
'rki': {'clean': 1, 'noisy': 1}, |
|
'rm': {'clean': 1, 'noisy': 1}, |
|
'rmc': {'clean': 1, 'noisy': 1}, |
|
'rn': {'clean': 1, 'noisy': 1}, |
|
'ro': {'clean': 20, 'noisy': 33}, |
|
'rom': {'clean': 1, 'noisy': 1}, |
|
'ru': {'clean': 299, 'noisy': 512}, |
|
'ru_Latn': {'clean': 1, 'noisy': 1}, |
|
'rw': {'clean': 1, 'noisy': 1}, |
|
'rwo': {'clean': 1, 'noisy': 1}, |
|
'sa': {'clean': 1, 'noisy': 1}, |
|
'sah': {'clean': 1, 'noisy': 1}, |
|
'sat_Latn': {'clean': 1, 'noisy': 1}, |
|
'sd': {'clean': 1, 'noisy': 1}, |
|
'sda': {'clean': 1, 'noisy': 1}, |
|
'se': {'clean': 1, 'noisy': 1}, |
|
'seh': {'clean': 1, 'noisy': 1}, |
|
'sg': {'clean': 1, 'noisy': 1}, |
|
'shn': {'clean': 1, 'noisy': 1}, |
|
'shp': {'clean': 1, 'noisy': 1}, |
|
'si': {'clean': 1, 'noisy': 1}, |
|
'sja': {'clean': 1, 'noisy': 1}, |
|
'sk': {'clean': 7, 'noisy': 12}, |
|
'skr': {'clean': 1, 'noisy': 1}, |
|
'sl': {'clean': 5, 'noisy': 7}, |
|
'sm': {'clean': 1, 'noisy': 1}, |
|
'smt': {'clean': 1, 'noisy': 1}, |
|
'sn': {'clean': 1, 'noisy': 2}, |
|
'so': {'clean': 1, 'noisy': 1}, |
|
'spp': {'clean': 1, 'noisy': 1}, |
|
'sq': {'clean': 2, 'noisy': 3}, |
|
'sr': {'clean': 2, 'noisy': 4}, |
|
'srm': {'clean': 1, 'noisy': 1}, |
|
'srn': {'clean': 1, 'noisy': 1}, |
|
'ss': {'clean': 1, 'noisy': 1}, |
|
'st': {'clean': 1, 'noisy': 1}, |
|
'stq': {'clean': 1, 'noisy': 1}, |
|
'su': {'clean': 1, 'noisy': 1}, |
|
'sus': {'clean': 1, 'noisy': 1}, |
|
'suz': {'clean': 1, 'noisy': 1}, |
|
'sv': {'clean': 21, 'noisy': 55}, |
|
'sw': {'clean': 1, 'noisy': 1}, |
|
'sxn': {'clean': 1, 'noisy': 1}, |
|
'syr': {'clean': 1, 'noisy': 1}, |
|
'ta': {'clean': 2, 'noisy': 4}, |
|
'ta_Latn': {'clean': 1, 'noisy': 1}, |
|
'tab': {'clean': 1, 'noisy': 1}, |
|
'taj': {'clean': 1, 'noisy': 1}, |
|
'tbz': {'clean': 1, 'noisy': 1}, |
|
'tca': {'clean': 1, 'noisy': 1}, |
|
'tcy': {'clean': 1, 'noisy': 1}, |
|
'tdx': {'clean': 1, 'noisy': 1}, |
|
'te': {'clean': 1, 'noisy': 2}, |
|
'te_Latn': {'clean': 1, 'noisy': 1}, |
|
'teo': {'clean': 1, 'noisy': 1}, |
|
'tet': {'clean': 1, 'noisy': 1}, |
|
'tg': {'clean': 1, 'noisy': 1}, |
|
'th': {'clean': 21, 'noisy': 22}, |
|
'ti': {'clean': 1, 'noisy': 1}, |
|
'tiv': {'clean': 1, 'noisy': 1}, |
|
'tk': {'clean': 1, 'noisy': 1}, |
|
'tks': {'clean': 1, 'noisy': 1}, |
|
'tlh': {'clean': 1, 'noisy': 1}, |
|
'tll': {'clean': 1, 'noisy': 1}, |
|
'tly_IR': {'clean': 1, 'noisy': 1}, |
|
'tn': {'clean': 1, 'noisy': 1}, |
|
'to': {'clean': 1, 'noisy': 1}, |
|
'toj': {'clean': 1, 'noisy': 1}, |
|
'tr': {'clean': 27, 'noisy': 50}, |
|
'trp': {'clean': 1, 'noisy': 1}, |
|
'ts': {'clean': 1, 'noisy': 1}, |
|
'tsc': {'clean': 1, 'noisy': 1}, |
|
'tsg': {'clean': 1, 'noisy': 1}, |
|
'tt': {'clean': 1, 'noisy': 2}, |
|
'tuc': {'clean': 1, 'noisy': 1}, |
|
'tvl': {'clean': 1, 'noisy': 1}, |
|
'twu': {'clean': 1, 'noisy': 1}, |
|
'tyv': {'clean': 1, 'noisy': 1}, |
|
'tyz': {'clean': 1, 'noisy': 1}, |
|
'tzh': {'clean': 1, 'noisy': 1}, |
|
'tzj': {'clean': 1, 'noisy': 1}, |
|
'tzo': {'clean': 1, 'noisy': 1}, |
|
'ubu': {'clean': 1, 'noisy': 1}, |
|
'udm': {'clean': 1, 'noisy': 1}, |
|
'ug': {'clean': 1, 'noisy': 1}, |
|
'uk': {'clean': 17, 'noisy': 28}, |
|
'ur': {'clean': 1, 'noisy': 1}, |
|
'uz': {'clean': 1, 'noisy': 1}, |
|
've': {'clean': 1, 'noisy': 1}, |
|
'vec': {'clean': 1, 'noisy': 1}, |
|
'vi': {'clean': 32, 'noisy': 49}, |
|
'wa': {'clean': 1, 'noisy': 1}, |
|
'wal': {'clean': 1, 'noisy': 1}, |
|
'war': {'clean': 1, 'noisy': 1}, |
|
'wo': {'clean': 1, 'noisy': 1}, |
|
'xal': {'clean': 1, 'noisy': 1}, |
|
'xh': {'clean': 1, 'noisy': 1}, |
|
'yap': {'clean': 1, 'noisy': 1}, |
|
'yi': {'clean': 1, 'noisy': 1}, |
|
'yo': {'clean': 1, 'noisy': 1}, |
|
'yua': {'clean': 1, 'noisy': 1}, |
|
'zap': {'clean': 1, 'noisy': 1}, |
|
'zh': {'clean': 40, 'noisy': 69}, |
|
'zh_Latn': {'clean': 1, 'noisy': 1}, |
|
'zne': {'clean': 1, 'noisy': 1}, |
|
'zu': {'clean': 1, 'noisy': 1}, |
|
'zxx_xx_dtynoise': {'clean': 1, 'noisy': 1}, |
|
'zza': {'clean': 1, 'noisy': 1} |
|
} |
|
|
|
_LANGUAGES = list(_N_SHARDS_PER_SPLIT.keys()) |
|
|
|
class Madlad400Config(datasets.BuilderConfig): |
|
"""BuilderConfig for madlad-400.""" |
|
|
|
def __init__(self, *args, languages, **kwargs): |
|
"""BuilderConfig for Madlad400. |
|
Args: |
|
languages (:obj:`List[str]`): list of languages to load |
|
**kwargs: keyword arguments forwarded to super. |
|
""" |
|
super().__init__( |
|
*args, |
|
name="+".join(languages), |
|
**kwargs, |
|
) |
|
self.languages = languages |
|
|
|
|
|
class Madlad400(datasets.GeneratorBasedBuilder): |
|
"""Madlad400,""" |
|
|
|
BUILDER_CONFIGS = [Madlad400Config(languages=[lang]) for lang in _LANGUAGES] |
|
BUILDER_CONFIG_CLASS = Madlad400Config |
|
|
|
def _info(self): |
|
return datasets.DatasetInfo( |
|
description=_DESCRIPTION, |
|
features=datasets.Features( |
|
{ |
|
"text": datasets.Value("string"), |
|
} |
|
), |
|
supervised_keys=None, |
|
citation=_CITATION, |
|
) |
|
|
|
def _split_generators(self, dl_manager): |
|
data_urls = {} |
|
for split in ["clean", "noisy"]: |
|
data_urls[split] = [ |
|
_DATA_URL.format( |
|
split=split, |
|
language=lang, |
|
index=index, |
|
) |
|
for lang in self.config.languages |
|
for index in range(_N_SHARDS_PER_SPLIT[lang][split]) |
|
] |
|
clean_downloaded_files = dl_manager.download(data_urls['clean']) |
|
noisy_downloaded_files = dl_manager.download(data_urls['noisy']) |
|
return [ |
|
datasets.SplitGenerator(name="clean", gen_kwargs={"filepaths": clean_downloaded_files}), |
|
datasets.SplitGenerator( |
|
name="noisy", gen_kwargs={"filepaths": noisy_downloaded_files} |
|
), |
|
] |
|
|
|
def _generate_examples(self, filepaths): |
|
"""This function returns the examples in the raw (text) form by iterating over the provided files.""" |
|
id_ = 0 |
|
for filepath in filepaths: |
|
logger.info("generating examples from = %s", filepath) |
|
with gzip.open(open(filepath, "rb"), "rt", encoding="utf-8") as f: |
|
for line in f: |
|
if line: |
|
example = json.loads(line) |
|
yield id_, example |
|
id_ += 1 |
|
|