donut-base-pdfs / tokenizer_config.json
savankamahina's picture
End of training
5187904 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_policy_information>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57526": {
"content": "</s_coverage_period>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57527": {
"content": "<s_policy_terms>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57528": {
"content": "<s_endorsements>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57529": {
"content": "</s_code>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57530": {
"content": "</s_nominee_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57531": {
"content": "</s_address>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57532": {
"content": "</s_name>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57533": {
"content": "<s_contact_number>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57534": {
"content": "<s_entity_type>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57535": {
"content": "</s_phone_number>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57536": {
"content": "<s_ncb_percentage>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57537": {
"content": "</s_policy class>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57538": {
"content": "<s_gst_number>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57539": {
"content": "<s_contact_info>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57540": {
"content": "</s_insurance_type>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57541": {
"content": "</s_coverage_type>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57542": {
"content": "<s_coverage_period>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57543": {
"content": "</s_ncb_percentage>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57544": {
"content": "</s_contact_number>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57545": {
"content": "</s_policy_number>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57546": {
"content": "<s_start_date>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57547": {
"content": "</s_insurer_name>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57548": {
"content": "<s_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57549": {
"content": "</s_end_date>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57550": {
"content": "<s_policy_number>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57551": {
"content": "</s_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57552": {
"content": "<s_code>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57553": {
"content": "</s_gst_number>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57554": {
"content": "</s_insured_entity>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57555": {
"content": "</s_total_premium>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57556": {
"content": "<s_additional_information>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57557": {
"content": "<s_phone_number>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57558": {
"content": "</s_dob>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57559": {
"content": "<s_registration_date>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57560": {
"content": "<s_coverage_type>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57561": {
"content": "<s_previous_policy_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57562": {
"content": "<s_policy_owner>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57563": {
"content": "<s_agent_broker_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57564": {
"content": "<s_exclusions>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57565": {
"content": "</s_entity_type>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57566": {
"content": "<s_nominee_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57567": {
"content": "<s_insurer_name>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57568": {
"content": "<s_dob>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57569": {
"content": "</s_contact_info>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57570": {
"content": "<s_make/model>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57571": {
"content": "<s_insurance_type>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57572": {
"content": "</s_agent_broker_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57573": {
"content": "<s_end_date>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57574": {
"content": "</s_endorsements>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57575": {
"content": "</s_registration_date>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57576": {
"content": "<s_policy_information>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57577": {
"content": "<s_address>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57578": {
"content": "<s_coverage_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57579": {
"content": "</s_email>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57580": {
"content": "<s_registration_no>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57581": {
"content": "</s_start_date>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57582": {
"content": "<s_name>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57583": {
"content": "<s_relationship>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57584": {
"content": "</s_insurance_provider>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57585": {
"content": "</s_exclusions>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57586": {
"content": "</s_previous_policy_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57587": {
"content": "</s_additional_information>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57588": {
"content": "<s_email>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57589": {
"content": "</s_policy_terms>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57590": {
"content": "<s_total_premium>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57591": {
"content": "</s_make/model>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57592": {
"content": "</s_coverage_details>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57593": {
"content": "</s_policy_owner>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57594": {
"content": "</s_relationship>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57595": {
"content": "<s_insured_entity>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57596": {
"content": "</s_registration_no>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57597": {
"content": "<s_insurance_provider>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"57598": {
"content": "<s_policy class>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"</s_policy_information>",
"</s_coverage_period>",
"<s_policy_terms>",
"<s_endorsements>",
"</s_code>",
"</s_nominee_details>",
"</s_address>",
"</s_name>",
"<s_contact_number>",
"<s_entity_type>",
"</s_phone_number>",
"<s_ncb_percentage>",
"</s_policy class>",
"<s_gst_number>",
"<s_contact_info>",
"</s_insurance_type>",
"</s_coverage_type>",
"<s_coverage_period>",
"</s_ncb_percentage>",
"</s_contact_number>",
"</s_policy_number>",
"<s_start_date>",
"</s_insurer_name>",
"<s_details>",
"</s_end_date>",
"<s_policy_number>",
"</s_details>",
"<s_code>",
"</s_gst_number>",
"</s_insured_entity>",
"</s_total_premium>",
"<s_additional_information>",
"<s_phone_number>",
"</s_dob>",
"<s_registration_date>",
"<s_coverage_type>",
"<s_previous_policy_details>",
"<s_policy_owner>",
"<s_agent_broker_details>",
"<s_exclusions>",
"<s>",
"</s_entity_type>",
"<s_nominee_details>",
"<s_insurer_name>",
"<s_dob>",
"</s_contact_info>",
"<s_make/model>",
"<s_insurance_type>",
"</s_agent_broker_details>",
"<s_end_date>",
"</s_endorsements>",
"</s_registration_date>",
"<s_policy_information>",
"<s_address>",
"<s_coverage_details>",
"</s_email>",
"<s_registration_no>",
"</s_start_date>",
"</s>",
"<s_name>",
"<s_relationship>",
"</s_insurance_provider>",
"</s_exclusions>",
"</s_previous_policy_details>",
"</s_additional_information>",
"<s_email>",
"</s_policy_terms>",
"<s_total_premium>",
"</s_make/model>",
"</s_coverage_details>",
"</s_policy_owner>",
"</s_relationship>",
"<s_insured_entity>",
"</s_registration_no>",
"<s_insurance_provider>",
"<s_policy class>"
],
"bos_token": "<s>",
"clean_up_tokenization_spaces": true,
"cls_token": "<s>",
"eos_token": "</s>",
"mask_token": "<mask>",
"max_length": 512,
"model_max_length": 1000000000000000019884624838656,
"pad_to_multiple_of": null,
"pad_token": "<pad>",
"pad_token_type_id": 0,
"padding_side": "right",
"processor_class": "DonutProcessor",
"sep_token": "</s>",
"sp_model_kwargs": {},
"stride": 0,
"tokenizer_class": "XLMRobertaTokenizer",
"truncation_side": "right",
"truncation_strategy": "longest_first",
"unk_token": "<unk>"
}