| { | |
| "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 | |
| }, | |
| "256001": { | |
| "content": "ace_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256002": { | |
| "content": "ace_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256003": { | |
| "content": "acm_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256004": { | |
| "content": "acq_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256005": { | |
| "content": "aeb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256006": { | |
| "content": "afr_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256007": { | |
| "content": "ajp_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256008": { | |
| "content": "aka_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256009": { | |
| "content": "amh_Ethi", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256010": { | |
| "content": "apc_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256011": { | |
| "content": "arb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256012": { | |
| "content": "ars_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256013": { | |
| "content": "ary_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256014": { | |
| "content": "arz_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256015": { | |
| "content": "asm_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256016": { | |
| "content": "ast_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256017": { | |
| "content": "awa_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256018": { | |
| "content": "ayr_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256019": { | |
| "content": "azb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256020": { | |
| "content": "azj_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256021": { | |
| "content": "bak_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256022": { | |
| "content": "bam_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256023": { | |
| "content": "ban_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256024": { | |
| "content": "bel_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256025": { | |
| "content": "bem_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256026": { | |
| "content": "ben_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256027": { | |
| "content": "bho_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256028": { | |
| "content": "bjn_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256029": { | |
| "content": "bjn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256030": { | |
| "content": "bod_Tibt", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256031": { | |
| "content": "bos_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256032": { | |
| "content": "bug_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256033": { | |
| "content": "bul_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256034": { | |
| "content": "cat_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256035": { | |
| "content": "ceb_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256036": { | |
| "content": "ces_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256037": { | |
| "content": "cjk_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256038": { | |
| "content": "ckb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256039": { | |
| "content": "crh_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256040": { | |
| "content": "cym_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256041": { | |
| "content": "dan_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256042": { | |
| "content": "deu_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256043": { | |
| "content": "dik_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256044": { | |
| "content": "dyu_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256045": { | |
| "content": "dzo_Tibt", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256046": { | |
| "content": "ell_Grek", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256047": { | |
| "content": "eng_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256048": { | |
| "content": "epo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256049": { | |
| "content": "est_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256050": { | |
| "content": "eus_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256051": { | |
| "content": "ewe_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256052": { | |
| "content": "fao_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256053": { | |
| "content": "pes_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256054": { | |
| "content": "fij_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256055": { | |
| "content": "fin_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256056": { | |
| "content": "fon_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256057": { | |
| "content": "fra_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256058": { | |
| "content": "fur_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256059": { | |
| "content": "fuv_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256060": { | |
| "content": "gla_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256061": { | |
| "content": "gle_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256062": { | |
| "content": "glg_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256063": { | |
| "content": "grn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256064": { | |
| "content": "guj_Gujr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256065": { | |
| "content": "hat_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256066": { | |
| "content": "hau_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256067": { | |
| "content": "heb_Hebr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256068": { | |
| "content": "hin_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256069": { | |
| "content": "hne_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256070": { | |
| "content": "hrv_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256071": { | |
| "content": "hun_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256072": { | |
| "content": "hye_Armn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256073": { | |
| "content": "ibo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256074": { | |
| "content": "ilo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256075": { | |
| "content": "ind_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256076": { | |
| "content": "isl_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256077": { | |
| "content": "ita_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256078": { | |
| "content": "jav_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256079": { | |
| "content": "jpn_Jpan", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256080": { | |
| "content": "kab_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256081": { | |
| "content": "kac_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256082": { | |
| "content": "kam_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256083": { | |
| "content": "kan_Knda", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256084": { | |
| "content": "kas_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256085": { | |
| "content": "kas_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256086": { | |
| "content": "kat_Geor", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256087": { | |
| "content": "knc_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256088": { | |
| "content": "knc_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256089": { | |
| "content": "kaz_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256090": { | |
| "content": "kbp_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256091": { | |
| "content": "kea_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256092": { | |
| "content": "khm_Khmr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256093": { | |
| "content": "kik_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256094": { | |
| "content": "kin_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256095": { | |
| "content": "kir_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256096": { | |
| "content": "kmb_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256097": { | |
| "content": "kon_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256098": { | |
| "content": "kor_Hang", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256099": { | |
| "content": "kmr_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256100": { | |
| "content": "lao_Laoo", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256101": { | |
| "content": "lvs_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256102": { | |
| "content": "lij_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256103": { | |
| "content": "lim_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256104": { | |
| "content": "lin_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256105": { | |
| "content": "lit_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256106": { | |
| "content": "lmo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256107": { | |
| "content": "ltg_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256108": { | |
| "content": "ltz_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256109": { | |
| "content": "lua_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256110": { | |
| "content": "lug_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256111": { | |
| "content": "luo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256112": { | |
| "content": "lus_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256113": { | |
| "content": "mag_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256114": { | |
| "content": "mai_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256115": { | |
| "content": "mal_Mlym", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256116": { | |
| "content": "mar_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256117": { | |
| "content": "min_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256118": { | |
| "content": "mkd_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256119": { | |
| "content": "plt_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256120": { | |
| "content": "mlt_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256121": { | |
| "content": "mni_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256122": { | |
| "content": "khk_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256123": { | |
| "content": "mos_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256124": { | |
| "content": "mri_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256125": { | |
| "content": "zsm_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256126": { | |
| "content": "mya_Mymr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256127": { | |
| "content": "nld_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256128": { | |
| "content": "nno_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256129": { | |
| "content": "nob_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256130": { | |
| "content": "npi_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256131": { | |
| "content": "nso_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256132": { | |
| "content": "nus_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256133": { | |
| "content": "nya_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256134": { | |
| "content": "oci_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256135": { | |
| "content": "gaz_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256136": { | |
| "content": "ory_Orya", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256137": { | |
| "content": "pag_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256138": { | |
| "content": "pan_Guru", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256139": { | |
| "content": "pap_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256140": { | |
| "content": "pol_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256141": { | |
| "content": "por_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256142": { | |
| "content": "prs_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256143": { | |
| "content": "pbt_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256144": { | |
| "content": "quy_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256145": { | |
| "content": "ron_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256146": { | |
| "content": "run_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256147": { | |
| "content": "rus_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256148": { | |
| "content": "sag_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256149": { | |
| "content": "san_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256150": { | |
| "content": "sat_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256151": { | |
| "content": "scn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256152": { | |
| "content": "shn_Mymr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256153": { | |
| "content": "sin_Sinh", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256154": { | |
| "content": "slk_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256155": { | |
| "content": "slv_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256156": { | |
| "content": "smo_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256157": { | |
| "content": "sna_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256158": { | |
| "content": "snd_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256159": { | |
| "content": "som_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256160": { | |
| "content": "sot_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256161": { | |
| "content": "spa_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256162": { | |
| "content": "als_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256163": { | |
| "content": "srd_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256164": { | |
| "content": "srp_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256165": { | |
| "content": "ssw_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256166": { | |
| "content": "sun_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256167": { | |
| "content": "swe_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256168": { | |
| "content": "swh_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256169": { | |
| "content": "szl_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256170": { | |
| "content": "tam_Taml", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256171": { | |
| "content": "tat_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256172": { | |
| "content": "tel_Telu", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256173": { | |
| "content": "tgk_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256174": { | |
| "content": "tgl_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256175": { | |
| "content": "tha_Thai", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256176": { | |
| "content": "tir_Ethi", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256177": { | |
| "content": "taq_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256178": { | |
| "content": "taq_Tfng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256179": { | |
| "content": "tpi_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256180": { | |
| "content": "tsn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256181": { | |
| "content": "tso_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256182": { | |
| "content": "tuk_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256183": { | |
| "content": "tum_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256184": { | |
| "content": "tur_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256185": { | |
| "content": "twi_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256186": { | |
| "content": "tzm_Tfng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256187": { | |
| "content": "uig_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256188": { | |
| "content": "ukr_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256189": { | |
| "content": "umb_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256190": { | |
| "content": "urd_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256191": { | |
| "content": "uzn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256192": { | |
| "content": "vec_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256193": { | |
| "content": "vie_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256194": { | |
| "content": "war_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256195": { | |
| "content": "wol_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256196": { | |
| "content": "xho_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256197": { | |
| "content": "ydd_Hebr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256198": { | |
| "content": "yor_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256199": { | |
| "content": "yue_Hant", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256200": { | |
| "content": "zho_Hans", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256201": { | |
| "content": "zho_Hant", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256202": { | |
| "content": "zul_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256203": { | |
| "content": "<mask>", | |
| "lstrip": true, | |
| "normalized": true, | |
| "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", | |
| "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", | |
| "pes_Arab", | |
| "fij_Latn", | |
| "fin_Latn", | |
| "fon_Latn", | |
| "fra_Latn", | |
| "fur_Latn", | |
| "fuv_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", | |
| "knc_Arab", | |
| "knc_Latn", | |
| "kaz_Cyrl", | |
| "kbp_Latn", | |
| "kea_Latn", | |
| "khm_Khmr", | |
| "kik_Latn", | |
| "kin_Latn", | |
| "kir_Cyrl", | |
| "kmb_Latn", | |
| "kon_Latn", | |
| "kor_Hang", | |
| "kmr_Latn", | |
| "lao_Laoo", | |
| "lvs_Latn", | |
| "lij_Latn", | |
| "lim_Latn", | |
| "lin_Latn", | |
| "lit_Latn", | |
| "lmo_Latn", | |
| "ltg_Latn", | |
| "ltz_Latn", | |
| "lua_Latn", | |
| "lug_Latn", | |
| "luo_Latn", | |
| "lus_Latn", | |
| "mag_Deva", | |
| "mai_Deva", | |
| "mal_Mlym", | |
| "mar_Deva", | |
| "min_Latn", | |
| "mkd_Cyrl", | |
| "plt_Latn", | |
| "mlt_Latn", | |
| "mni_Beng", | |
| "khk_Cyrl", | |
| "mos_Latn", | |
| "mri_Latn", | |
| "zsm_Latn", | |
| "mya_Mymr", | |
| "nld_Latn", | |
| "nno_Latn", | |
| "nob_Latn", | |
| "npi_Deva", | |
| "nso_Latn", | |
| "nus_Latn", | |
| "nya_Latn", | |
| "oci_Latn", | |
| "gaz_Latn", | |
| "ory_Orya", | |
| "pag_Latn", | |
| "pan_Guru", | |
| "pap_Latn", | |
| "pol_Latn", | |
| "por_Latn", | |
| "prs_Arab", | |
| "pbt_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", | |
| "als_Latn", | |
| "srd_Latn", | |
| "srp_Cyrl", | |
| "ssw_Latn", | |
| "sun_Latn", | |
| "swe_Latn", | |
| "swh_Latn", | |
| "szl_Latn", | |
| "tam_Taml", | |
| "tat_Cyrl", | |
| "tel_Telu", | |
| "tgk_Cyrl", | |
| "tgl_Latn", | |
| "tha_Thai", | |
| "tir_Ethi", | |
| "taq_Latn", | |
| "taq_Tfng", | |
| "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", | |
| "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": "khm_Khmr", | |
| "tgt_lang": "eng_Latn", | |
| "tokenizer_class": "NllbTokenizer", | |
| "unk_token": "<unk>" | |
| } | |