| { | |
| "added_tokens_decoder": { | |
| "0": { | |
| "content": "<s>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "1": { | |
| "content": "<pad>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "2": { | |
| "content": "</s>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "3": { | |
| "content": "<unk>", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262209": { | |
| "content": "<mask>", | |
| "lstrip": true, | |
| "normalized": true, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262210": { | |
| "content": "ace_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262211": { | |
| "content": "ace_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262212": { | |
| "content": "acm_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262213": { | |
| "content": "acq_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262214": { | |
| "content": "aeb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262215": { | |
| "content": "afr_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262216": { | |
| "content": "ajp_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262217": { | |
| "content": "aka_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262218": { | |
| "content": "als_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262219": { | |
| "content": "amh_Ethi", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262220": { | |
| "content": "apc_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262221": { | |
| "content": "arb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262222": { | |
| "content": "ars_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262223": { | |
| "content": "ary_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262224": { | |
| "content": "arz_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262225": { | |
| "content": "asm_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262226": { | |
| "content": "ast_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262227": { | |
| "content": "awa_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262228": { | |
| "content": "ayr_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262229": { | |
| "content": "azb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262230": { | |
| "content": "azj_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262231": { | |
| "content": "bak_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262232": { | |
| "content": "bam_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262233": { | |
| "content": "ban_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262234": { | |
| "content": "bel_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262235": { | |
| "content": "bem_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262236": { | |
| "content": "ben_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262237": { | |
| "content": "bho_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262238": { | |
| "content": "bjn_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262239": { | |
| "content": "bjn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262240": { | |
| "content": "bod_Tibt", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262241": { | |
| "content": "bos_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262242": { | |
| "content": "bug_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262243": { | |
| "content": "bul_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262244": { | |
| "content": "cat_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262245": { | |
| "content": "ceb_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262246": { | |
| "content": "ces_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262247": { | |
| "content": "cjk_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262248": { | |
| "content": "ckb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262249": { | |
| "content": "crh_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262250": { | |
| "content": "cym_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262251": { | |
| "content": "dan_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262252": { | |
| "content": "deu_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262253": { | |
| "content": "dik_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262254": { | |
| "content": "dyu_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262255": { | |
| "content": "dzo_Tibt", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262256": { | |
| "content": "ell_Grek", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262257": { | |
| "content": "eng_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262258": { | |
| "content": "epo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262259": { | |
| "content": "est_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262260": { | |
| "content": "eus_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262261": { | |
| "content": "ewe_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262262": { | |
| "content": "fao_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262263": { | |
| "content": "fij_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262264": { | |
| "content": "fin_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262265": { | |
| "content": "fon_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262266": { | |
| "content": "fra_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262267": { | |
| "content": "fur_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262268": { | |
| "content": "fuv_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262269": { | |
| "content": "gaz_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262270": { | |
| "content": "gla_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262271": { | |
| "content": "gle_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262272": { | |
| "content": "glg_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262273": { | |
| "content": "grn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262274": { | |
| "content": "guj_Gujr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262275": { | |
| "content": "hat_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262276": { | |
| "content": "hau_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262277": { | |
| "content": "heb_Hebr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262278": { | |
| "content": "hin_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262279": { | |
| "content": "hne_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262280": { | |
| "content": "hrv_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262281": { | |
| "content": "hun_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262282": { | |
| "content": "hye_Armn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262283": { | |
| "content": "ibo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262284": { | |
| "content": "ilo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262285": { | |
| "content": "ind_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262286": { | |
| "content": "isl_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262287": { | |
| "content": "ita_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262288": { | |
| "content": "jav_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262289": { | |
| "content": "jpn_Jpan", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262290": { | |
| "content": "kab_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262291": { | |
| "content": "kac_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262292": { | |
| "content": "kam_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262293": { | |
| "content": "kan_Knda", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262294": { | |
| "content": "kas_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262295": { | |
| "content": "kas_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262296": { | |
| "content": "kat_Geor", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262297": { | |
| "content": "kaz_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262298": { | |
| "content": "kbp_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262299": { | |
| "content": "kea_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262300": { | |
| "content": "khk_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262301": { | |
| "content": "khm_Khmr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262302": { | |
| "content": "kik_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262303": { | |
| "content": "kin_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262304": { | |
| "content": "kir_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262305": { | |
| "content": "kmb_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262306": { | |
| "content": "kmr_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262307": { | |
| "content": "knc_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262308": { | |
| "content": "knc_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262309": { | |
| "content": "kon_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262310": { | |
| "content": "kor_Hang", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262311": { | |
| "content": "lao_Laoo", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262312": { | |
| "content": "lij_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262313": { | |
| "content": "lim_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262314": { | |
| "content": "lin_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262315": { | |
| "content": "lit_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262316": { | |
| "content": "lmo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262317": { | |
| "content": "ltg_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262318": { | |
| "content": "ltz_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262319": { | |
| "content": "lua_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262320": { | |
| "content": "lug_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262321": { | |
| "content": "luo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262322": { | |
| "content": "lus_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262323": { | |
| "content": "lvs_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262324": { | |
| "content": "mag_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262325": { | |
| "content": "mai_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262326": { | |
| "content": "mal_Mlym", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262327": { | |
| "content": "mar_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262328": { | |
| "content": "min_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262329": { | |
| "content": "mkd_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262330": { | |
| "content": "mlt_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262331": { | |
| "content": "mni_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262332": { | |
| "content": "mos_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262333": { | |
| "content": "mri_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262334": { | |
| "content": "mya_Mymr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262335": { | |
| "content": "myv_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262336": { | |
| "content": "nld_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262337": { | |
| "content": "nno_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262338": { | |
| "content": "nob_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262339": { | |
| "content": "npi_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262340": { | |
| "content": "nso_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262341": { | |
| "content": "nus_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262342": { | |
| "content": "nya_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262343": { | |
| "content": "oci_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262344": { | |
| "content": "ory_Orya", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262345": { | |
| "content": "pag_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262346": { | |
| "content": "pan_Guru", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262347": { | |
| "content": "pap_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262348": { | |
| "content": "pbt_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262349": { | |
| "content": "pes_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262350": { | |
| "content": "plt_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262351": { | |
| "content": "pol_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262352": { | |
| "content": "por_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262353": { | |
| "content": "prs_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262354": { | |
| "content": "quy_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262355": { | |
| "content": "ron_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262356": { | |
| "content": "run_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262357": { | |
| "content": "rus_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262358": { | |
| "content": "sag_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262359": { | |
| "content": "san_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262360": { | |
| "content": "sat_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262361": { | |
| "content": "scn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262362": { | |
| "content": "shn_Mymr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262363": { | |
| "content": "sin_Sinh", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262364": { | |
| "content": "slk_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262365": { | |
| "content": "slv_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262366": { | |
| "content": "smo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262367": { | |
| "content": "sna_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262368": { | |
| "content": "snd_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262369": { | |
| "content": "som_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262370": { | |
| "content": "sot_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262371": { | |
| "content": "spa_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262372": { | |
| "content": "srd_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262373": { | |
| "content": "srp_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262374": { | |
| "content": "ssw_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262375": { | |
| "content": "sun_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262376": { | |
| "content": "swe_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262377": { | |
| "content": "swh_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262378": { | |
| "content": "szl_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262379": { | |
| "content": "tam_Taml", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262380": { | |
| "content": "taq_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262381": { | |
| "content": "taq_Tfng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262382": { | |
| "content": "tat_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262383": { | |
| "content": "tel_Telu", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262384": { | |
| "content": "tgk_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262385": { | |
| "content": "tgl_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262386": { | |
| "content": "tha_Thai", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262387": { | |
| "content": "tir_Ethi", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262388": { | |
| "content": "tpi_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262389": { | |
| "content": "tsn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262390": { | |
| "content": "tso_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262391": { | |
| "content": "tuk_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262392": { | |
| "content": "tum_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262393": { | |
| "content": "tur_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262394": { | |
| "content": "twi_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262395": { | |
| "content": "tzm_Tfng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262396": { | |
| "content": "uig_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262397": { | |
| "content": "ukr_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262398": { | |
| "content": "umb_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262399": { | |
| "content": "urd_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262400": { | |
| "content": "uzn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262401": { | |
| "content": "vec_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262402": { | |
| "content": "vie_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262403": { | |
| "content": "war_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262404": { | |
| "content": "wol_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262405": { | |
| "content": "xho_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262406": { | |
| "content": "ydd_Hebr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262407": { | |
| "content": "yor_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262408": { | |
| "content": "yue_Hant", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262409": { | |
| "content": "zho_Hans", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262410": { | |
| "content": "zho_Hant", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262411": { | |
| "content": "zsm_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "262412": { | |
| "content": "zul_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| } | |
| }, | |
| "additional_special_tokens": [ | |
| "ace_Arab", | |
| "ace_Latn", | |
| "acm_Arab", | |
| "acq_Arab", | |
| "aeb_Arab", | |
| "afr_Latn", | |
| "ajp_Arab", | |
| "aka_Latn", | |
| "als_Latn", | |
| "amh_Ethi", | |
| "apc_Arab", | |
| "arb_Arab", | |
| "ars_Arab", | |
| "ary_Arab", | |
| "arz_Arab", | |
| "asm_Beng", | |
| "ast_Latn", | |
| "awa_Deva", | |
| "ayr_Latn", | |
| "azb_Arab", | |
| "azj_Latn", | |
| "bak_Cyrl", | |
| "bam_Latn", | |
| "ban_Latn", | |
| "bel_Cyrl", | |
| "bem_Latn", | |
| "ben_Beng", | |
| "bho_Deva", | |
| "bjn_Arab", | |
| "bjn_Latn", | |
| "bod_Tibt", | |
| "bos_Latn", | |
| "bug_Latn", | |
| "bul_Cyrl", | |
| "cat_Latn", | |
| "ceb_Latn", | |
| "ces_Latn", | |
| "cjk_Latn", | |
| "ckb_Arab", | |
| "crh_Latn", | |
| "cym_Latn", | |
| "dan_Latn", | |
| "deu_Latn", | |
| "dik_Latn", | |
| "dyu_Latn", | |
| "dzo_Tibt", | |
| "ell_Grek", | |
| "eng_Latn", | |
| "epo_Latn", | |
| "est_Latn", | |
| "eus_Latn", | |
| "ewe_Latn", | |
| "fao_Latn", | |
| "fij_Latn", | |
| "fin_Latn", | |
| "fon_Latn", | |
| "fra_Latn", | |
| "fur_Latn", | |
| "fuv_Latn", | |
| "gaz_Latn", | |
| "gla_Latn", | |
| "gle_Latn", | |
| "glg_Latn", | |
| "grn_Latn", | |
| "guj_Gujr", | |
| "hat_Latn", | |
| "hau_Latn", | |
| "heb_Hebr", | |
| "hin_Deva", | |
| "hne_Deva", | |
| "hrv_Latn", | |
| "hun_Latn", | |
| "hye_Armn", | |
| "ibo_Latn", | |
| "ilo_Latn", | |
| "ind_Latn", | |
| "isl_Latn", | |
| "ita_Latn", | |
| "jav_Latn", | |
| "jpn_Jpan", | |
| "kab_Latn", | |
| "kac_Latn", | |
| "kam_Latn", | |
| "kan_Knda", | |
| "kas_Arab", | |
| "kas_Deva", | |
| "kat_Geor", | |
| "kaz_Cyrl", | |
| "kbp_Latn", | |
| "kea_Latn", | |
| "khk_Cyrl", | |
| "khm_Khmr", | |
| "kik_Latn", | |
| "kin_Latn", | |
| "kir_Cyrl", | |
| "kmb_Latn", | |
| "kmr_Latn", | |
| "knc_Arab", | |
| "knc_Latn", | |
| "kon_Latn", | |
| "kor_Hang", | |
| "lao_Laoo", | |
| "lij_Latn", | |
| "lim_Latn", | |
| "lin_Latn", | |
| "lit_Latn", | |
| "lmo_Latn", | |
| "ltg_Latn", | |
| "ltz_Latn", | |
| "lua_Latn", | |
| "lug_Latn", | |
| "luo_Latn", | |
| "lus_Latn", | |
| "lvs_Latn", | |
| "mag_Deva", | |
| "mai_Deva", | |
| "mal_Mlym", | |
| "mar_Deva", | |
| "min_Latn", | |
| "mkd_Cyrl", | |
| "mlt_Latn", | |
| "mni_Beng", | |
| "mos_Latn", | |
| "mri_Latn", | |
| "mya_Mymr", | |
| "myv_Cyrl", | |
| "nld_Latn", | |
| "nno_Latn", | |
| "nob_Latn", | |
| "npi_Deva", | |
| "nso_Latn", | |
| "nus_Latn", | |
| "nya_Latn", | |
| "oci_Latn", | |
| "ory_Orya", | |
| "pag_Latn", | |
| "pan_Guru", | |
| "pap_Latn", | |
| "pbt_Arab", | |
| "pes_Arab", | |
| "plt_Latn", | |
| "pol_Latn", | |
| "por_Latn", | |
| "prs_Arab", | |
| "quy_Latn", | |
| "ron_Latn", | |
| "run_Latn", | |
| "rus_Cyrl", | |
| "sag_Latn", | |
| "san_Deva", | |
| "sat_Beng", | |
| "scn_Latn", | |
| "shn_Mymr", | |
| "sin_Sinh", | |
| "slk_Latn", | |
| "slv_Latn", | |
| "smo_Latn", | |
| "sna_Latn", | |
| "snd_Arab", | |
| "som_Latn", | |
| "sot_Latn", | |
| "spa_Latn", | |
| "srd_Latn", | |
| "srp_Cyrl", | |
| "ssw_Latn", | |
| "sun_Latn", | |
| "swe_Latn", | |
| "swh_Latn", | |
| "szl_Latn", | |
| "tam_Taml", | |
| "taq_Latn", | |
| "taq_Tfng", | |
| "tat_Cyrl", | |
| "tel_Telu", | |
| "tgk_Cyrl", | |
| "tgl_Latn", | |
| "tha_Thai", | |
| "tir_Ethi", | |
| "tpi_Latn", | |
| "tsn_Latn", | |
| "tso_Latn", | |
| "tuk_Latn", | |
| "tum_Latn", | |
| "tur_Latn", | |
| "twi_Latn", | |
| "tzm_Tfng", | |
| "uig_Arab", | |
| "ukr_Cyrl", | |
| "umb_Latn", | |
| "urd_Arab", | |
| "uzn_Latn", | |
| "vec_Latn", | |
| "vie_Latn", | |
| "war_Latn", | |
| "wol_Latn", | |
| "xho_Latn", | |
| "ydd_Hebr", | |
| "yor_Latn", | |
| "yue_Hant", | |
| "zho_Hans", | |
| "zho_Hant", | |
| "zsm_Latn", | |
| "zul_Latn" | |
| ], | |
| "bos_token": "<s>", | |
| "clean_up_tokenization_spaces": true, | |
| "cls_token": "<s>", | |
| "eos_token": "</s>", | |
| "legacy_behaviour": false, | |
| "mask_token": "<mask>", | |
| "model_max_length": 1024, | |
| "pad_token": "<pad>", | |
| "sep_token": "</s>", | |
| "sp_model_kwargs": {}, | |
| "src_lang": "myv_Cyrl", | |
| "tgt_lang": "rus_Cyrl", | |
| "tokenizer_class": "NllbTokenizer", | |
| "unk_token": "<unk>" | |
| } | |