| { | |
| "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": "hin_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256002": { | |
| "content": "ben_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256003": { | |
| "content": "mar_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256004": { | |
| "content": "tel_Telu", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256005": { | |
| "content": "tam_Taml", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256006": { | |
| "content": "guj_Gujr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256007": { | |
| "content": "kan_Knda", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256008": { | |
| "content": "mal_Mlym", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256009": { | |
| "content": "asm_Beng", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256010": { | |
| "content": "urd_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256011": { | |
| "content": "ory_Orya", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256012": { | |
| "content": "eng_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256013": { | |
| "content": "fra_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256014": { | |
| "content": "ita_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256015": { | |
| "content": "spa_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256016": { | |
| "content": "deu_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256017": { | |
| "content": "por_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256018": { | |
| "content": "kor_Hang", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256019": { | |
| "content": "tha_Thai", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256020": { | |
| "content": "arb_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256021": { | |
| "content": "vie_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256022": { | |
| "content": "rus_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256023": { | |
| "content": "jpn_Jpan", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256024": { | |
| "content": "tur_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256025": { | |
| "content": "ukr_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256026": { | |
| "content": "pes_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256027": { | |
| "content": "nld_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256028": { | |
| "content": "npi_Deva", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256029": { | |
| "content": "pol_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256030": { | |
| "content": "ind_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256031": { | |
| "content": "afr_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256032": { | |
| "content": "amh_Ethi", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256033": { | |
| "content": "hye_Armn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256034": { | |
| "content": "bam_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256035": { | |
| "content": "bel_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256036": { | |
| "content": "bul_Cyrl", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256037": { | |
| "content": "zho_Hans", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256038": { | |
| "content": "ces_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256039": { | |
| "content": "dan_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256040": { | |
| "content": "fin_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256041": { | |
| "content": "ell_Grek", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256042": { | |
| "content": "heb_Hebr", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256043": { | |
| "content": "mri_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256044": { | |
| "content": "ron_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256045": { | |
| "content": "slv_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256046": { | |
| "content": "swe_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256047": { | |
| "content": "xho_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256048": { | |
| "content": "uig_Arab", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256049": { | |
| "content": "som_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256050": { | |
| "content": "pan_Guru", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256051": { | |
| "content": "hin_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256052": { | |
| "content": "ben_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256053": { | |
| "content": "tel_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256054": { | |
| "content": "tam_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256055": { | |
| "content": "guj_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256056": { | |
| "content": "kan_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256057": { | |
| "content": "arb_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256058": { | |
| "content": "rus_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256059": { | |
| "content": "jpn_Latn", | |
| "lstrip": false, | |
| "normalized": false, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| }, | |
| "256060": { | |
| "content": "<mask>", | |
| "lstrip": false, | |
| "normalized": true, | |
| "rstrip": false, | |
| "single_word": false, | |
| "special": true | |
| } | |
| }, | |
| "additional_special_tokens": [ | |
| "hin_Deva", | |
| "ben_Beng", | |
| "mar_Deva", | |
| "tel_Telu", | |
| "tam_Taml", | |
| "guj_Gujr", | |
| "kan_Knda", | |
| "mal_Mlym", | |
| "asm_Beng", | |
| "urd_Arab", | |
| "ory_Orya", | |
| "eng_Latn", | |
| "fra_Latn", | |
| "ita_Latn", | |
| "spa_Latn", | |
| "deu_Latn", | |
| "por_Latn", | |
| "kor_Hang", | |
| "tha_Thai", | |
| "arb_Arab", | |
| "vie_Latn", | |
| "rus_Cyrl", | |
| "jpn_Jpan", | |
| "tur_Latn", | |
| "ukr_Cyrl", | |
| "pes_Arab", | |
| "nld_Latn", | |
| "npi_Deva", | |
| "pol_Latn", | |
| "ind_Latn", | |
| "afr_Latn", | |
| "amh_Ethi", | |
| "hye_Armn", | |
| "bam_Latn", | |
| "bel_Cyrl", | |
| "bul_Cyrl", | |
| "zho_Hans", | |
| "ces_Latn", | |
| "dan_Latn", | |
| "fin_Latn", | |
| "ell_Grek", | |
| "heb_Hebr", | |
| "mri_Latn", | |
| "ron_Latn", | |
| "slv_Latn", | |
| "swe_Latn", | |
| "xho_Latn", | |
| "uig_Arab", | |
| "som_Latn", | |
| "pan_Guru", | |
| "hin_Latn", | |
| "ben_Latn", | |
| "tel_Latn", | |
| "tam_Latn", | |
| "guj_Latn", | |
| "kan_Latn", | |
| "arb_Latn", | |
| "rus_Latn", | |
| "jpn_Latn" | |
| ], | |
| "bos_token": "<s>", | |
| "clean_up_tokenization_spaces": true, | |
| "cls_token": "<s>", | |
| "eos_token": "</s>", | |
| "mask_token": "<mask>", | |
| "model_max_length": 1000000000000000019884624838656, | |
| "pad_token": "<pad>", | |
| "sep_token": "</s>", | |
| "tokenizer_class": "PreTrainedTokenizerFast", | |
| "unk_token": "<unk>" | |
| } | |