donut-finetune-20sample-LB / tokenizer_config.json
revathyds31's picture
Upload processor
a3e3066
{
"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_answer>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57524": {
"content": "</s_question>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57525": {
"content": "<no/>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57526": {
"content": "<s_answer>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57527": {
"content": "<s_docvqa>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57528": {
"content": "<s_iitcdip>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57529": {
"content": "<s_question>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57530": {
"content": "<s_synthdog>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57531": {
"content": "<yes/>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57532": {
"content": "<s_pres_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57533": {
"content": "</s_pres_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57534": {
"content": "<s_pres_fax>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57535": {
"content": "</s_pres_fax>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57536": {
"content": "<s_pati_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57537": {
"content": "</s_pati_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57538": {
"content": "<s_pati_dob>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57539": {
"content": "</s_pati_dob>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57540": {
"content": "<s_pati_eoc>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57541": {
"content": "</s_pati_eoc>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57542": {
"content": "<s_drug_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57543": {
"content": "</s_drug_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57544": {
"content": "<s_j_code>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57545": {
"content": "</s_j_code>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57546": {
"content": "<s_pati_address>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57547": {
"content": "</s_pati_address>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57548": {
"content": "<s_pati_phone>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57549": {
"content": "</s_pati_phone>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57550": {
"content": "<s_pati_member_id>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57551": {
"content": "</s_pati_member_id>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57552": {
"content": "<s_pati_zip>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57553": {
"content": "</s_pati_zip>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57554": {
"content": "<s_pati_group_number>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57555": {
"content": "</s_pati_group_number>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57556": {
"content": "<s_pati_city_state_zip>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57557": {
"content": "</s_pati_city_state_zip>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57558": {
"content": "<s_pres_tax_id>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57559": {
"content": "</s_pres_tax_id>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57560": {
"content": "<s_pres_address>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57561": {
"content": "</s_pres_address>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57562": {
"content": "<s_pres_state_lic_id>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57563": {
"content": "</s_pres_state_lic_id>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57564": {
"content": "<s_pres_phone>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57565": {
"content": "</s_pres_phone>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57566": {
"content": "<s_pres_npi>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57567": {
"content": "</s_pres_npi>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57568": {
"content": "<s_pres_city_state_zip>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57569": {
"content": "</s_pres_city_state_zip>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57570": {
"content": "<s_direction_sig>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57571": {
"content": "</s_direction_sig>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57572": {
"content": "<s_drug_quantity>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57573": {
"content": "</s_drug_quantity>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57574": {
"content": "<s_prov_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57575": {
"content": "</s_prov_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57576": {
"content": "<s_prov_npi>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57577": {
"content": "</s_prov_npi>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57578": {
"content": "<s_prov_address>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57579": {
"content": "</s_prov_address>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57580": {
"content": "<s_prov_fax>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57581": {
"content": "</s_prov_fax>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57582": {
"content": "<s_prov_phone>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"57583": {
"content": "</s_prov_phone>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
}
},
"additional_special_tokens": [
"</s_answer>",
"</s_question>",
"<no/>",
"<s_answer>",
"<s_docvqa>",
"<s_iitcdip>",
"<s_question>",
"<s_synthdog>",
"<yes/>"
],
"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>"
}