{ "added_tokens_decoder": { "0": { "content": "", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "1": { "content": "", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "2": { "content": "", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "3": { "content": "", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282723": { "content": "", "lstrip": true, "normalized": true, "rstrip": false, "single_word": false, "special": true }, "282724": { "content": "ace_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282725": { "content": "ace_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282726": { "content": "acm_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282727": { "content": "acq_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282728": { "content": "aeb_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282729": { "content": "afr_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282730": { "content": "ajp_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282731": { "content": "aka_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282732": { "content": "als_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282733": { "content": "amh_Ethi", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282734": { "content": "apc_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282735": { "content": "arb_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282736": { "content": "ars_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282737": { "content": "ary_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282738": { "content": "arz_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282739": { "content": "asm_Beng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282740": { "content": "ast_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282741": { "content": "awa_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282742": { "content": "ayr_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282743": { "content": "azb_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282744": { "content": "azj_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282745": { "content": "bak_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282746": { "content": "bam_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282747": { "content": "ban_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282748": { "content": "bel_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282749": { "content": "bem_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282750": { "content": "ben_Beng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282751": { "content": "bho_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282752": { "content": "bjn_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282753": { "content": "bjn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282754": { "content": "bod_Tibt", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282755": { "content": "bos_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282756": { "content": "bug_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282757": { "content": "bul_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282758": { "content": "bxr_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282759": { "content": "cat_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282760": { "content": "ceb_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282761": { "content": "ces_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282762": { "content": "che_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282763": { "content": "cjk_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282764": { "content": "ckb_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282765": { "content": "crh_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282766": { "content": "cym_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282767": { "content": "dan_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282768": { "content": "deu_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282769": { "content": "dik_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282770": { "content": "dyu_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282771": { "content": "dzo_Tibt", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282772": { "content": "ell_Grek", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282773": { "content": "eng_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282774": { "content": "epo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282775": { "content": "est_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282776": { "content": "eus_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282777": { "content": "ewe_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282778": { "content": "fao_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282779": { "content": "fij_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282780": { "content": "fin_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282781": { "content": "fon_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282782": { "content": "fra_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282783": { "content": "fur_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282784": { "content": "fuv_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282785": { "content": "gaz_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282786": { "content": "gla_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282787": { "content": "gle_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282788": { "content": "glg_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282789": { "content": "grn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282790": { "content": "guj_Gujr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282791": { "content": "hat_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282792": { "content": "hau_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282793": { "content": "heb_Hebr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282794": { "content": "hin_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282795": { "content": "hne_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282796": { "content": "hrv_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282797": { "content": "hun_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282798": { "content": "hye_Armn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282799": { "content": "ibo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282800": { "content": "ilo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282801": { "content": "ind_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282802": { "content": "isl_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282803": { "content": "ita_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282804": { "content": "jav_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282805": { "content": "jpn_Jpan", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282806": { "content": "kab_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282807": { "content": "kac_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282808": { "content": "kam_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282809": { "content": "kan_Knda", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282810": { "content": "kas_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282811": { "content": "kas_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282812": { "content": "kat_Geor", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282813": { "content": "kaz_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282814": { "content": "kbp_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282815": { "content": "kea_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282816": { "content": "khk_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282817": { "content": "khm_Khmr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282818": { "content": "kik_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282819": { "content": "kin_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282820": { "content": "kir_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282821": { "content": "kmb_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282822": { "content": "kmr_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282823": { "content": "knc_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282824": { "content": "knc_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282825": { "content": "kon_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282826": { "content": "kor_Hang", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282827": { "content": "lao_Laoo", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282828": { "content": "lij_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282829": { "content": "lim_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282830": { "content": "lin_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282831": { "content": "lit_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282832": { "content": "lmo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282833": { "content": "ltg_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282834": { "content": "ltz_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282835": { "content": "lua_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282836": { "content": "lug_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282837": { "content": "luo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282838": { "content": "lus_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282839": { "content": "lvs_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282840": { "content": "mag_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282841": { "content": "mai_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282842": { "content": "mal_Mlym", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282843": { "content": "mar_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282844": { "content": "min_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282845": { "content": "mkd_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282846": { "content": "mlt_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282847": { "content": "mni_Beng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282848": { "content": "mos_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282849": { "content": "mri_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282850": { "content": "mya_Mymr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282851": { "content": "myv_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282852": { "content": "nld_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282853": { "content": "nno_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282854": { "content": "nob_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282855": { "content": "npi_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282856": { "content": "nso_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282857": { "content": "nus_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282858": { "content": "nya_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282859": { "content": "oci_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282860": { "content": "ory_Orya", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282861": { "content": "pag_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282862": { "content": "pan_Guru", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282863": { "content": "pap_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282864": { "content": "pbt_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282865": { "content": "pes_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282866": { "content": "plt_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282867": { "content": "pol_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282868": { "content": "por_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282869": { "content": "prs_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282870": { "content": "quy_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282871": { "content": "ron_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282872": { "content": "run_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282873": { "content": "rus_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282874": { "content": "sag_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282875": { "content": "san_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282876": { "content": "sat_Beng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282877": { "content": "scn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282878": { "content": "shn_Mymr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282879": { "content": "sin_Sinh", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282880": { "content": "slk_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282881": { "content": "slv_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282882": { "content": "smo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282883": { "content": "sna_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282884": { "content": "snd_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282885": { "content": "som_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282886": { "content": "sot_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282887": { "content": "spa_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282888": { "content": "srd_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282889": { "content": "srp_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282890": { "content": "ssw_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282891": { "content": "sun_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282892": { "content": "swe_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282893": { "content": "swh_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282894": { "content": "szl_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282895": { "content": "tam_Taml", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282896": { "content": "taq_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282897": { "content": "taq_Tfng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282898": { "content": "tat_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282899": { "content": "tel_Telu", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282900": { "content": "tgk_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282901": { "content": "tgl_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282902": { "content": "tha_Thai", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282903": { "content": "tir_Ethi", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282904": { "content": "tpi_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282905": { "content": "tsn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282906": { "content": "tso_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282907": { "content": "tuk_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282908": { "content": "tum_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282909": { "content": "tur_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282910": { "content": "twi_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282911": { "content": "tyv_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282912": { "content": "tzm_Tfng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282913": { "content": "uig_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282914": { "content": "ukr_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282915": { "content": "umb_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282916": { "content": "urd_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282917": { "content": "uzn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282918": { "content": "vec_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282919": { "content": "vie_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282920": { "content": "war_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282921": { "content": "wol_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282922": { "content": "xho_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282923": { "content": "ydd_Hebr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282924": { "content": "yor_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282925": { "content": "yue_Hant", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282926": { "content": "zho_Hans", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282927": { "content": "zho_Hant", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282928": { "content": "zsm_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "282929": { "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", "bxr_Cyrl", "cat_Latn", "ceb_Latn", "ces_Latn", "che_Cyrl", "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", "tyv_Cyrl", "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": "", "clean_up_tokenization_spaces": true, "cls_token": "", "eos_token": "", "legacy_behaviour": false, "mask_token": "", "model_max_length": 1024, "pad_token": "", "sep_token": "", "sp_model_kwargs": {}, "src_lang": "myv_Cyrl", "tgt_lang": "rus_Cyrl", "tokenizer_class": "NllbTokenizer", "unk_token": "" }