|
{ |
|
"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_VLF_CLASS>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57526": { |
|
"content": "</s_RESISTERED_OWNER>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57527": { |
|
"content": "</s_REGISTRATION_VALID_FROM>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57528": { |
|
"content": "</s_PIC>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57529": { |
|
"content": "<s_LIENHOLDER>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57530": { |
|
"content": "<s_UNLADEN/G/CGW>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57531": { |
|
"content": "<s_BODY_TYPE_MODEL>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57532": { |
|
"content": "<s_USE_TAX>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57533": { |
|
"content": "</s_AMOUNT_PAID>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57534": { |
|
"content": "</s_PR_EXP_DATE>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57535": { |
|
"content": "<s_MP>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57536": { |
|
"content": "<s_TYPE_VEHICLE_USE>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57537": { |
|
"content": "</s_DT_FEE_RECVD>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57538": { |
|
"content": "<s_MAKE>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57539": { |
|
"content": "<s_YR_star>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57540": { |
|
"content": "<s_VLF_CLASS>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57541": { |
|
"content": "</s_UNLADEN/G/CGW>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57542": { |
|
"content": "<s_PR_EXP_DATE>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57543": { |
|
"content": "</s_TYPE>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57544": { |
|
"content": "</s_WC>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57545": { |
|
"content": "<s_TYPE>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57546": { |
|
"content": "</s_YR_MODEL>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57547": { |
|
"content": "<s_PIC>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57548": { |
|
"content": "<s_REGISTRATION_VALID_FROM>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57549": { |
|
"content": "</s_AX>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57550": { |
|
"content": "</s_YR_star>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57551": { |
|
"content": "</s_TYPE_VEHICLE_USE>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57552": { |
|
"content": "<s_YR_1ST_SOLD>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57553": { |
|
"content": "</s_DATE_ISSUED>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57554": { |
|
"content": "</s_CC/ALCO>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57555": { |
|
"content": "</s_REGISTRATION_VALID_TO>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57556": { |
|
"content": "<s_CYLS>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57557": { |
|
"content": "<s_AX>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57558": { |
|
"content": "<s_DATE_ISSUED>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57559": { |
|
"content": "<s_VIN>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57560": { |
|
"content": "</s_MP>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57561": { |
|
"content": "</s_TYPE_VEH>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57562": { |
|
"content": "<s_YR_MODEL>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57563": { |
|
"content": "</s_CYLS>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57564": { |
|
"content": "</s_USE_TAX>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57565": { |
|
"content": "<s_MO>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57566": { |
|
"content": "<s_TYPE_VEH>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57567": { |
|
"content": "</s_LIENHOLDER>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57568": { |
|
"content": "</s_YR_1ST_SOLD>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57569": { |
|
"content": "<s_LICENSE_NUMBER>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57570": { |
|
"content": "</s_BODY_TYPE_MODEL>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57571": { |
|
"content": "</s_VIN>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57572": { |
|
"content": "<s_CC/ALCO>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57573": { |
|
"content": "<s_DT_FEE_RECVD>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57574": { |
|
"content": "<s_REGISTRATION_VALID_TO>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57575": { |
|
"content": "</s_MAKE>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57576": { |
|
"content": "</s_LICENSE_NUMBER>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57577": { |
|
"content": "<s_AMOUNT_PAID>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57578": { |
|
"content": "<s_RESISTERED_OWNER>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57579": { |
|
"content": "</s_MO>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57580": { |
|
"content": "<s_WC>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
} |
|
}, |
|
"additional_special_tokens": [ |
|
"</s_VLF_CLASS>", |
|
"</s_RESISTERED_OWNER>", |
|
"</s_REGISTRATION_VALID_FROM>", |
|
"</s_PIC>", |
|
"<s_LIENHOLDER>", |
|
"<s_UNLADEN/G/CGW>", |
|
"<s_BODY_TYPE_MODEL>", |
|
"<s_USE_TAX>", |
|
"</s_AMOUNT_PAID>", |
|
"</s_PR_EXP_DATE>", |
|
"<s_MP>", |
|
"<s_TYPE_VEHICLE_USE>", |
|
"</s_DT_FEE_RECVD>", |
|
"<s_MAKE>", |
|
"<s_YR_star>", |
|
"<s_VLF_CLASS>", |
|
"</s_UNLADEN/G/CGW>", |
|
"<s_PR_EXP_DATE>", |
|
"</s_TYPE>", |
|
"</s_WC>", |
|
"<s_TYPE>", |
|
"</s_YR_MODEL>", |
|
"<s_PIC>", |
|
"<s_REGISTRATION_VALID_FROM>", |
|
"</s_AX>", |
|
"</s_YR_star>", |
|
"</s_TYPE_VEHICLE_USE>", |
|
"<s>", |
|
"<s_YR_1ST_SOLD>", |
|
"</s_DATE_ISSUED>", |
|
"</s_CC/ALCO>", |
|
"</s_REGISTRATION_VALID_TO>", |
|
"<s_CYLS>", |
|
"<s_AX>", |
|
"<s_DATE_ISSUED>", |
|
"<s_VIN>", |
|
"</s_MP>", |
|
"</s_TYPE_VEH>", |
|
"<s_YR_MODEL>", |
|
"</s_CYLS>", |
|
"</s_USE_TAX>", |
|
"<s_MO>", |
|
"<s_TYPE_VEH>", |
|
"</s_LIENHOLDER>", |
|
"</s_YR_1ST_SOLD>", |
|
"<s_LICENSE_NUMBER>", |
|
"</s_BODY_TYPE_MODEL>", |
|
"</s>", |
|
"</s_VIN>", |
|
"<s_CC/ALCO>", |
|
"<s_DT_FEE_RECVD>", |
|
"<s_REGISTRATION_VALID_TO>", |
|
"</s_MAKE>", |
|
"</s_LICENSE_NUMBER>", |
|
"<s_AMOUNT_PAID>", |
|
"<s_RESISTERED_OWNER>", |
|
"</s_MO>", |
|
"<s_WC>" |
|
], |
|
"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>" |
|
} |
|
|