|
{ |
|
"added_tokens_decoder": { |
|
"0": { |
|
"content": "<s>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"1": { |
|
"content": "<pad>", |
|
"lstrip": false, |
|
"normalized": false, |
|
"rstrip": false, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"2": { |
|
"content": "</s>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"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": true, |
|
"normalized": true, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": false |
|
}, |
|
"57523": { |
|
"content": "<s_iitcdip>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57524": { |
|
"content": "<s_synthdog>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57525": { |
|
"content": "</s_gross_weight>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57526": { |
|
"content": "</s_purchase_order>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57527": { |
|
"content": "</s_item_index>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57528": { |
|
"content": "</s_HS_code>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57529": { |
|
"content": "</s_vendor>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57530": { |
|
"content": "</s_part_number>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57531": { |
|
"content": "</s_total_item_value>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57532": { |
|
"content": "</s_item_quantity>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57533": { |
|
"content": "</s_line>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57534": { |
|
"content": "<s_part_number>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57535": { |
|
"content": "</s_item_number>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57536": { |
|
"content": "<s_HS_code>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57537": { |
|
"content": "<s_line>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57538": { |
|
"content": "</s_origin_country>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57539": { |
|
"content": "<s_hs_code>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57540": { |
|
"content": "<s_item_value>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57541": { |
|
"content": "</s_hs_code>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57542": { |
|
"content": "<s_net_weight>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57543": { |
|
"content": "<s_item_quantity>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57544": { |
|
"content": "<s_item_number>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57545": { |
|
"content": "</s_description>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57546": { |
|
"content": "<s_origin_country>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57547": { |
|
"content": "<s_discount>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57548": { |
|
"content": "<s_gross_weight>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57549": { |
|
"content": "</s_net_weight>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57550": { |
|
"content": "</s_item_value>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57551": { |
|
"content": "<s_uom>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57552": { |
|
"content": "<s_manf_part_number>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57553": { |
|
"content": "</s_uom>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57554": { |
|
"content": "</s_manf_part_number>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57555": { |
|
"content": "<s_delivery_number>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57556": { |
|
"content": "</s_delivery_number>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57557": { |
|
"content": "<s_quantity_type>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57558": { |
|
"content": "<s_purchase_order>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57559": { |
|
"content": "<s_description>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57560": { |
|
"content": "<s_total_item_value>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57561": { |
|
"content": "</s_quantity_type>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57562": { |
|
"content": "</s_discount>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57563": { |
|
"content": "<s_vendor>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
}, |
|
"57564": { |
|
"content": "<s_item_index>", |
|
"lstrip": true, |
|
"normalized": false, |
|
"rstrip": true, |
|
"single_word": false, |
|
"special": true |
|
} |
|
}, |
|
"additional_special_tokens": [ |
|
"<s_iitcdip>", |
|
"<s_synthdog>", |
|
"</s_gross_weight>", |
|
"</s_purchase_order>", |
|
"</s_item_index>", |
|
"</s_HS_code>", |
|
"</s_vendor>", |
|
"</s_part_number>", |
|
"</s_total_item_value>", |
|
"</s_item_quantity>", |
|
"</s_line>", |
|
"<s_part_number>", |
|
"</s_item_number>", |
|
"<s_HS_code>", |
|
"<s_line>", |
|
"</s_origin_country>", |
|
"<s_hs_code>", |
|
"<s_item_value>", |
|
"</s_hs_code>", |
|
"<s_net_weight>", |
|
"<s_item_quantity>", |
|
"<s_item_number>", |
|
"</s_description>", |
|
"<s>", |
|
"<s_origin_country>", |
|
"<s_discount>", |
|
"<s_gross_weight>", |
|
"</s_net_weight>", |
|
"</s_item_value>", |
|
"<s_uom>", |
|
"<s_manf_part_number>", |
|
"</s_uom>", |
|
"</s_manf_part_number>", |
|
"<s_delivery_number>", |
|
"</s_delivery_number>", |
|
"<s_quantity_type>", |
|
"<s_purchase_order>", |
|
"<s_description>", |
|
"</s>", |
|
"<s_total_item_value>", |
|
"</s_quantity_type>", |
|
"</s_discount>", |
|
"<s_vendor>", |
|
"<s_item_index>" |
|
], |
|
"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>" |
|
} |
|
|