donut-base-sroie / tokenizer_config.json
Saksham278's picture
End of training
e6bb989 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
},
"57521": {
"content": "<mask>",
"lstrip": true,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": true
},
"57522": {
"content": "<sep/>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57523": {
"content": "<s_iitcdip>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57524": {
"content": "<s_synthdog>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57525": {
"content": "<s_STATE_LIC_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57526": {
"content": "</s_STATE_LIC_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57527": {
"content": "<s_STATE_LIC_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57528": {
"content": "</s_STATE_LIC_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57529": {
"content": "<s_STATE_LIC_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57530": {
"content": "</s_STATE_LIC_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57531": {
"content": "<s_STATE_LIC_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57532": {
"content": "</s_STATE_LIC_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57533": {
"content": "<s_STATE_LIC_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57534": {
"content": "</s_STATE_LIC_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57535": {
"content": "<s_SEX_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57536": {
"content": "</s_SEX_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57537": {
"content": "<s_SEX_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57538": {
"content": "</s_SEX_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57539": {
"content": "<s_SEX_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57540": {
"content": "</s_SEX_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57541": {
"content": "<s_SEX_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57542": {
"content": "</s_SEX_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57543": {
"content": "<s_SEX_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57544": {
"content": "</s_SEX_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57545": {
"content": "<s_NAME_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57546": {
"content": "</s_NAME_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57547": {
"content": "<s_NAME_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57548": {
"content": "</s_NAME_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57549": {
"content": "<s_NAME_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57550": {
"content": "</s_NAME_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57551": {
"content": "<s_NAME_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57552": {
"content": "</s_NAME_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57553": {
"content": "<s_NAME_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57554": {
"content": "</s_NAME_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57555": {
"content": "<s_MAR_STAT_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57556": {
"content": "</s_MAR_STAT_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57557": {
"content": "<s_MAR_STAT_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57558": {
"content": "</s_MAR_STAT_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57559": {
"content": "<s_MAR_STAT_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57560": {
"content": "</s_MAR_STAT_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57561": {
"content": "<s_MAR_STAT_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57562": {
"content": "</s_MAR_STAT_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57563": {
"content": "<s_MAR_STAT_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57564": {
"content": "</s_MAR_STAT_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57565": {
"content": "<s_DRIVING_LICENCE_NUMBER_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57566": {
"content": "</s_DRIVING_LICENCE_NUMBER_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57567": {
"content": "<s_DRIVING_LICENCE_NUMBER_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57568": {
"content": "</s_DRIVING_LICENCE_NUMBER_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57569": {
"content": "<s_DRIVING_LICENCE_NUMBER_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57570": {
"content": "</s_DRIVING_LICENCE_NUMBER_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57571": {
"content": "<s_DRIVING_LICENCE_NUMBER_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57572": {
"content": "</s_DRIVING_LICENCE_NUMBER_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57573": {
"content": "<s_DRIVING_LICENCE_NUMBER_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57574": {
"content": "</s_DRIVING_LICENCE_NUMBER_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57575": {
"content": "<s_DRIVER_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57576": {
"content": "</s_DRIVER_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57577": {
"content": "<s_DRIVER_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57578": {
"content": "</s_DRIVER_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57579": {
"content": "<s_DRIVER_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57580": {
"content": "</s_DRIVER_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57581": {
"content": "<s_DRIVER_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57582": {
"content": "</s_DRIVER_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57583": {
"content": "<s_DRIVER_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57584": {
"content": "</s_DRIVER_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57585": {
"content": "<s_DATE_OF_BIRTH_formated_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57586": {
"content": "</s_DATE_OF_BIRTH_formated_5>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57587": {
"content": "<s_DATE_OF_BIRTH_formated_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57588": {
"content": "</s_DATE_OF_BIRTH_formated_4>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57589": {
"content": "<s_DATE_OF_BIRTH_formated_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57590": {
"content": "</s_DATE_OF_BIRTH_formated_3>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57591": {
"content": "<s_DATE_OF_BIRTH_formated_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57592": {
"content": "</s_DATE_OF_BIRTH_formated_2>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57593": {
"content": "<s_DATE_OF_BIRTH_formated_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57594": {
"content": "</s_DATE_OF_BIRTH_formated_1>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"<s_STATE_LIC_5>",
"</s_STATE_LIC_5>",
"<s_STATE_LIC_4>",
"</s_STATE_LIC_4>",
"<s_STATE_LIC_3>",
"</s_STATE_LIC_3>",
"<s_STATE_LIC_2>",
"</s_STATE_LIC_2>",
"<s_STATE_LIC_1>",
"</s_STATE_LIC_1>",
"<s_SEX_5>",
"</s_SEX_5>",
"<s_SEX_4>",
"</s_SEX_4>",
"<s_SEX_3>",
"</s_SEX_3>",
"<s_SEX_2>",
"</s_SEX_2>",
"<s_SEX_1>",
"</s_SEX_1>",
"<s_NAME_5>",
"</s_NAME_5>",
"<s_NAME_4>",
"</s_NAME_4>",
"<s_NAME_3>",
"</s_NAME_3>",
"<s_NAME_2>",
"</s_NAME_2>",
"<s_NAME_1>",
"</s_NAME_1>",
"<s_MAR_STAT_5>",
"</s_MAR_STAT_5>",
"<s_MAR_STAT_4>",
"</s_MAR_STAT_4>",
"<s_MAR_STAT_3>",
"</s_MAR_STAT_3>",
"<s_MAR_STAT_2>",
"</s_MAR_STAT_2>",
"<s_MAR_STAT_1>",
"</s_MAR_STAT_1>",
"<s_DRIVING_LICENCE_NUMBER_5>",
"</s_DRIVING_LICENCE_NUMBER_5>",
"<s_DRIVING_LICENCE_NUMBER_4>",
"</s_DRIVING_LICENCE_NUMBER_4>",
"<s_DRIVING_LICENCE_NUMBER_3>",
"</s_DRIVING_LICENCE_NUMBER_3>",
"<s_DRIVING_LICENCE_NUMBER_2>",
"</s_DRIVING_LICENCE_NUMBER_2>",
"<s_DRIVING_LICENCE_NUMBER_1>",
"</s_DRIVING_LICENCE_NUMBER_1>",
"<s_DRIVER_5>",
"</s_DRIVER_5>",
"<s_DRIVER_4>",
"</s_DRIVER_4>",
"<s_DRIVER_3>",
"</s_DRIVER_3>",
"<s_DRIVER_2>",
"</s_DRIVER_2>",
"<s_DRIVER_1>",
"</s_DRIVER_1>",
"<s_DATE_OF_BIRTH_formated_5>",
"</s_DATE_OF_BIRTH_formated_5>",
"<s_DATE_OF_BIRTH_formated_4>",
"</s_DATE_OF_BIRTH_formated_4>",
"<s_DATE_OF_BIRTH_formated_3>",
"</s_DATE_OF_BIRTH_formated_3>",
"<s_DATE_OF_BIRTH_formated_2>",
"</s_DATE_OF_BIRTH_formated_2>",
"<s_DATE_OF_BIRTH_formated_1>",
"</s_DATE_OF_BIRTH_formated_1>",
"<s>",
"</s>"
],
"bos_token": "<s>",
"clean_up_tokenization_spaces": true,
"cls_token": "<s>",
"eos_token": "</s>",
"mask_token": "<mask>",
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<pad>",
"processor_class": "DonutProcessor",
"sep_token": "</s>",
"sp_model_kwargs": {},
"tokenizer_class": "XLMRobertaTokenizer",
"unk_token": "<unk>"
}