mt5-budget-seq2seq-2 / tokenizer_config.json
napatswift's picture
Training in progress, step 100
22b903a
{
"added_tokens_decoder": {
"0": {
"content": "<pad>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"1": {
"content": "</s>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"2": {
"content": "<unk>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"31865": {
"content": "<root>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31866": {
"content": "</root>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31867": {
"content": "</amount>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31868": {
"content": "</fiscal_year>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31869": {
"content": "</fiscal_year_end_year>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31870": {
"content": "</fiscal_year_year>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31871": {
"content": "</line_item_amount>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31872": {
"content": "</line_item_bullet>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31873": {
"content": "</line_item_children>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31874": {
"content": "</line_item_fiscal_years>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31875": {
"content": "</line_item_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31876": {
"content": "</line_item_root>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31877": {
"content": "<amount>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31878": {
"content": "<begin>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31879": {
"content": "<begin_indent>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31880": {
"content": "<big_sep>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31881": {
"content": "<fiscal_year>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31882": {
"content": "<fiscal_year_end_year>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31883": {
"content": "<fiscal_year_year>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31884": {
"content": "<line_item_amount>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31885": {
"content": "<line_item_bullet>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31886": {
"content": "<line_item_children>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31887": {
"content": "<line_item_fiscal_years>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31888": {
"content": "<line_item_name>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
},
"31889": {
"content": "<line_item_root>",
"lstrip": false,
"normalized": true,
"rstrip": false,
"single_word": false,
"special": false
}
},
"additional_special_tokens": [
"<pad>",
"</s>",
"<unk>"
],
"clean_up_tokenization_spaces": true,
"eos_token": "</s>",
"extra_ids": 0,
"legacy": true,
"max_length": 1100,
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<pad>",
"sp_model_kwargs": {},
"stride": 0,
"tokenizer_class": "T5Tokenizer",
"truncation_side": "right",
"truncation_strategy": "longest_first",
"unk_token": "<unk>"
}