fine-tune-mBart-ar-en / tokenizer_config.json
Messam174's picture
Upload tokenizer
c39a338 verified
{
"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
},
"250001": {
"content": "ar_AR",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250002": {
"content": "cs_CZ",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250003": {
"content": "de_DE",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250004": {
"content": "en_XX",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250005": {
"content": "es_XX",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250006": {
"content": "et_EE",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250007": {
"content": "fi_FI",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250008": {
"content": "fr_XX",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250009": {
"content": "gu_IN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250010": {
"content": "hi_IN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250011": {
"content": "it_IT",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250012": {
"content": "ja_XX",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250013": {
"content": "kk_KZ",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250014": {
"content": "ko_KR",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250015": {
"content": "lt_LT",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250016": {
"content": "lv_LV",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250017": {
"content": "my_MM",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250018": {
"content": "ne_NP",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250019": {
"content": "nl_XX",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250020": {
"content": "ro_RO",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250021": {
"content": "ru_RU",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250022": {
"content": "si_LK",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250023": {
"content": "tr_TR",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250024": {
"content": "vi_VN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250025": {
"content": "zh_CN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250026": {
"content": "af_ZA",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250027": {
"content": "az_AZ",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250028": {
"content": "bn_IN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250029": {
"content": "fa_IR",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250030": {
"content": "he_IL",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250031": {
"content": "hr_HR",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250032": {
"content": "id_ID",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250033": {
"content": "ka_GE",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250034": {
"content": "km_KH",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250035": {
"content": "mk_MK",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250036": {
"content": "ml_IN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250037": {
"content": "mn_MN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250038": {
"content": "mr_IN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250039": {
"content": "pl_PL",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250040": {
"content": "ps_AF",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250041": {
"content": "pt_XX",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250042": {
"content": "sv_SE",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250043": {
"content": "sw_KE",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250044": {
"content": "ta_IN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250045": {
"content": "te_IN",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250046": {
"content": "th_TH",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250047": {
"content": "tl_XX",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250048": {
"content": "uk_UA",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250049": {
"content": "ur_PK",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250050": {
"content": "xh_ZA",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250051": {
"content": "gl_ES",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250052": {
"content": "sl_SI",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"250053": {
"content": "<mask>",
"lstrip": true,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"ar_AR",
"cs_CZ",
"de_DE",
"en_XX",
"es_XX",
"et_EE",
"fi_FI",
"fr_XX",
"gu_IN",
"hi_IN",
"it_IT",
"ja_XX",
"kk_KZ",
"ko_KR",
"lt_LT",
"lv_LV",
"my_MM",
"ne_NP",
"nl_XX",
"ro_RO",
"ru_RU",
"si_LK",
"tr_TR",
"vi_VN",
"zh_CN",
"af_ZA",
"az_AZ",
"bn_IN",
"fa_IR",
"he_IL",
"hr_HR",
"id_ID",
"ka_GE",
"km_KH",
"mk_MK",
"ml_IN",
"mn_MN",
"mr_IN",
"pl_PL",
"ps_AF",
"pt_XX",
"sv_SE",
"sw_KE",
"ta_IN",
"te_IN",
"th_TH",
"tl_XX",
"uk_UA",
"ur_PK",
"xh_ZA",
"gl_ES",
"sl_SI"
],
"bos_token": "<s>",
"clean_up_tokenization_spaces": false,
"cls_token": "<s>",
"eos_token": "</s>",
"language_codes": "ML50",
"mask_token": "<mask>",
"max_length": 128,
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<pad>",
"return_tensors": "pt",
"sep_token": "</s>",
"sp_model_kwargs": {},
"src_lang": "en_XX",
"stride": 0,
"tgt_lang": "ar_AR",
"tokenizer_class": "MBart50Tokenizer",
"truncation_side": "right",
"truncation_strategy": "longest_first",
"unk_token": "<unk>"
}