eacortes commited on
Commit
d776e32
·
verified ·
1 Parent(s): e13bb88

Upload 2 files

Browse files
Files changed (1) hide show
  1. tokenizer_config.json +7 -0
tokenizer_config.json CHANGED
@@ -45,13 +45,20 @@
45
  "cls_token": "[CLS]",
46
  "extra_special_tokens": {},
47
  "mask_token": "[MASK]",
 
48
  "model_input_names": [
49
  "input_ids",
50
  "attention_mask"
51
  ],
52
  "model_max_length": 512,
 
53
  "pad_token": "[PAD]",
 
 
54
  "sep_token": "[SEP]",
 
55
  "tokenizer_class": "PreTrainedTokenizerFast",
 
 
56
  "unk_token": "[UNK]"
57
  }
 
45
  "cls_token": "[CLS]",
46
  "extra_special_tokens": {},
47
  "mask_token": "[MASK]",
48
+ "max_length": 512,
49
  "model_input_names": [
50
  "input_ids",
51
  "attention_mask"
52
  ],
53
  "model_max_length": 512,
54
+ "pad_to_multiple_of": null,
55
  "pad_token": "[PAD]",
56
+ "pad_token_type_id": 0,
57
+ "padding_side": "right",
58
  "sep_token": "[SEP]",
59
+ "stride": 0,
60
  "tokenizer_class": "PreTrainedTokenizerFast",
61
+ "truncation_side": "right",
62
+ "truncation_strategy": "longest_first",
63
  "unk_token": "[UNK]"
64
  }