{ "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 }, "395270": { "content": "", "lstrip": true, "normalized": true, "rstrip": false, "single_word": false, "special": true }, "395271": { "content": "ace_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395272": { "content": "ace_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395273": { "content": "acm_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395274": { "content": "acq_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395275": { "content": "aeb_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395276": { "content": "afr_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395277": { "content": "ajp_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395278": { "content": "aka_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395279": { "content": "als_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395280": { "content": "amh_Ethi", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395281": { "content": "apc_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395282": { "content": "arb_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395283": { "content": "ars_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395284": { "content": "ary_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395285": { "content": "arz_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395286": { "content": "asm_Beng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395287": { "content": "ast_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395288": { "content": "awa_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395289": { "content": "ayr_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395290": { "content": "azb_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395291": { "content": "azj_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395292": { "content": "bak_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395293": { "content": "bam_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395294": { "content": "ban_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395295": { "content": "bel_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395296": { "content": "bem_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395297": { "content": "ben_Beng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395298": { "content": "bho_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395299": { "content": "bjn_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395300": { "content": "bjn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395301": { "content": "bod_Tibt", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395302": { "content": "bos_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395303": { "content": "bug_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395304": { "content": "bul_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395305": { "content": "bxr_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395306": { "content": "cat_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395307": { "content": "ceb_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395308": { "content": "ces_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395309": { "content": "che_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395310": { "content": "chv_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395311": { "content": "cjk_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395312": { "content": "ckb_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395313": { "content": "crh_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395314": { "content": "cym_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395315": { "content": "dan_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395316": { "content": "deu_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395317": { "content": "dik_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395318": { "content": "dyu_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395319": { "content": "dzo_Tibt", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395320": { "content": "ell_Grek", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395321": { "content": "eng_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395322": { "content": "epo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395323": { "content": "est_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395324": { "content": "eus_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395325": { "content": "ewe_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395326": { "content": "fao_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395327": { "content": "fij_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395328": { "content": "fin_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395329": { "content": "fon_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395330": { "content": "fra_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395331": { "content": "fur_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395332": { "content": "fuv_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395333": { "content": "gaz_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395334": { "content": "gla_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395335": { "content": "gle_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395336": { "content": "glg_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395337": { "content": "grn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395338": { "content": "guj_Gujr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395339": { "content": "hat_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395340": { "content": "hau_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395341": { "content": "heb_Hebr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395342": { "content": "hin_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395343": { "content": "hne_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395344": { "content": "hrv_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395345": { "content": "hun_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395346": { "content": "hye_Armn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395347": { "content": "ibo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395348": { "content": "ilo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395349": { "content": "ind_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395350": { "content": "isl_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395351": { "content": "ita_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395352": { "content": "jav_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395353": { "content": "jpn_Jpan", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395354": { "content": "kab_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395355": { "content": "kac_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395356": { "content": "kam_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395357": { "content": "kan_Knda", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395358": { "content": "kas_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395359": { "content": "kas_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395360": { "content": "kat_Geor", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395361": { "content": "kaz_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395362": { "content": "kbp_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395363": { "content": "kea_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395364": { "content": "khk_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395365": { "content": "khm_Khmr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395366": { "content": "kik_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395367": { "content": "kin_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395368": { "content": "kir_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395369": { "content": "kjh_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395370": { "content": "kmb_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395371": { "content": "kmr_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395372": { "content": "knc_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395373": { "content": "knc_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395374": { "content": "kon_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395375": { "content": "kor_Hang", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395376": { "content": "krc_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395377": { "content": "lao_Laoo", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395378": { "content": "lij_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395379": { "content": "lim_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395380": { "content": "lin_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395381": { "content": "lit_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395382": { "content": "lmo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395383": { "content": "ltg_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395384": { "content": "ltz_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395385": { "content": "lua_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395386": { "content": "lug_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395387": { "content": "luo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395388": { "content": "lus_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395389": { "content": "lvs_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395390": { "content": "mag_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395391": { "content": "mai_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395392": { "content": "mal_Mlym", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395393": { "content": "mar_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395394": { "content": "mhr_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395395": { "content": "min_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395396": { "content": "mkd_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395397": { "content": "mlt_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395398": { "content": "mni_Beng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395399": { "content": "mos_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395400": { "content": "mri_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395401": { "content": "mya_Mymr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395402": { "content": "myv_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395403": { "content": "nld_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395404": { "content": "nno_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395405": { "content": "nob_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395406": { "content": "npi_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395407": { "content": "nso_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395408": { "content": "nus_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395409": { "content": "nya_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395410": { "content": "oci_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395411": { "content": "ory_Orya", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395412": { "content": "pag_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395413": { "content": "pan_Guru", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395414": { "content": "pap_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395415": { "content": "pbt_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395416": { "content": "pes_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395417": { "content": "plt_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395418": { "content": "pol_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395419": { "content": "por_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395420": { "content": "prs_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395421": { "content": "quy_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395422": { "content": "ron_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395423": { "content": "run_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395424": { "content": "rus_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395425": { "content": "sag_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395426": { "content": "san_Deva", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395427": { "content": "sat_Beng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395428": { "content": "scn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395429": { "content": "shn_Mymr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395430": { "content": "sin_Sinh", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395431": { "content": "slk_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395432": { "content": "slv_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395433": { "content": "smo_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395434": { "content": "sna_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395435": { "content": "snd_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395436": { "content": "som_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395437": { "content": "sot_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395438": { "content": "spa_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395439": { "content": "srd_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395440": { "content": "srp_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395441": { "content": "ssw_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395442": { "content": "sun_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395443": { "content": "swe_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395444": { "content": "swh_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395445": { "content": "szl_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395446": { "content": "tam_Taml", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395447": { "content": "taq_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395448": { "content": "taq_Tfng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395449": { "content": "tat_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395450": { "content": "tel_Telu", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395451": { "content": "tgk_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395452": { "content": "tgl_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395453": { "content": "tha_Thai", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395454": { "content": "tir_Ethi", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395455": { "content": "tpi_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395456": { "content": "tsn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395457": { "content": "tso_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395458": { "content": "tuk_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395459": { "content": "tum_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395460": { "content": "tur_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395461": { "content": "twi_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395462": { "content": "tyv_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395463": { "content": "tzm_Tfng", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395464": { "content": "uig_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395465": { "content": "ukr_Cyrl", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395466": { "content": "umb_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395467": { "content": "urd_Arab", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395468": { "content": "uzn_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395469": { "content": "vec_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395470": { "content": "vie_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395471": { "content": "war_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395472": { "content": "wol_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395473": { "content": "xho_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395474": { "content": "ydd_Hebr", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395475": { "content": "yor_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395476": { "content": "yue_Hant", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395477": { "content": "zho_Hans", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395478": { "content": "zho_Hant", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395479": { "content": "zsm_Latn", "lstrip": false, "normalized": false, "rstrip": false, "single_word": false, "special": true }, "395480": { "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", "chv_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", "kjh_Cyrl", "kmb_Latn", "kmr_Latn", "knc_Arab", "knc_Latn", "kon_Latn", "kor_Hang", "krc_Cyrl", "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", "mhr_Cyrl", "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": "rus_Cyrl", "tgt_lang": "bak_Cyrl", "tokenizer_class": "NllbTokenizer", "unk_token": "" }