TeeA commited on
Commit
9ad70a5
1 Parent(s): 12f7854

Upload processor

Browse files
Files changed (2) hide show
  1. special_tokens_map.json +21 -3
  2. tokenizer_config.json +8 -0
special_tokens_map.json CHANGED
@@ -97,7 +97,25 @@
97
  "<extra_id_94>",
98
  "<extra_id_95>"
99
  ],
100
- "eos_token": "</s>",
101
- "pad_token": "<pad>",
102
- "unk_token": "<unk>"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
103
  }
 
97
  "<extra_id_94>",
98
  "<extra_id_95>"
99
  ],
100
+ "eos_token": {
101
+ "content": "</s>",
102
+ "lstrip": false,
103
+ "normalized": false,
104
+ "rstrip": false,
105
+ "single_word": false
106
+ },
107
+ "pad_token": {
108
+ "content": "<pad>",
109
+ "lstrip": false,
110
+ "normalized": false,
111
+ "rstrip": false,
112
+ "single_word": false
113
+ },
114
+ "unk_token": {
115
+ "content": "<unk>",
116
+ "lstrip": false,
117
+ "normalized": false,
118
+ "rstrip": false,
119
+ "single_word": false
120
+ }
121
  }
tokenizer_config.json CHANGED
@@ -894,10 +894,18 @@
894
  "clean_up_tokenization_spaces": true,
895
  "eos_token": "</s>",
896
  "extra_ids": 96,
 
897
  "model_max_length": 1000000000000000019884624838656,
 
898
  "pad_token": "<pad>",
 
 
899
  "processor_class": "Pix2StructProcessor",
900
  "sp_model_kwargs": {},
 
901
  "tokenizer_class": "T5Tokenizer",
 
 
 
902
  "unk_token": "<unk>"
903
  }
 
894
  "clean_up_tokenization_spaces": true,
895
  "eos_token": "</s>",
896
  "extra_ids": 96,
897
+ "max_length": 256,
898
  "model_max_length": 1000000000000000019884624838656,
899
+ "pad_to_multiple_of": null,
900
  "pad_token": "<pad>",
901
+ "pad_token_type_id": 0,
902
+ "padding_side": "right",
903
  "processor_class": "Pix2StructProcessor",
904
  "sp_model_kwargs": {},
905
+ "stride": 0,
906
  "tokenizer_class": "T5Tokenizer",
907
+ "truncation_side": "right",
908
+ "truncation_strategy": "longest_first",
909
+ "trust_remote_code": false,
910
  "unk_token": "<unk>"
911
  }