UnderstandLing commited on
Commit
ba37be5
1 Parent(s): 3ee1b4f

Upload tokenizer

Browse files
Files changed (2) hide show
  1. special_tokens_map.json +1 -1
  2. tokenizer_config.json +2 -2
special_tokens_map.json CHANGED
@@ -14,7 +14,7 @@
14
  "single_word": false
15
  },
16
  "pad_token": {
17
- "content": "<unk>",
18
  "lstrip": false,
19
  "normalized": false,
20
  "rstrip": false,
 
14
  "single_word": false
15
  },
16
  "pad_token": {
17
+ "content": "<|endoftext|>",
18
  "lstrip": false,
19
  "normalized": false,
20
  "rstrip": false,
tokenizer_config.json CHANGED
@@ -122,8 +122,8 @@
122
  "legacy": false,
123
  "max_length": 512,
124
  "model_max_length": 131072,
125
- "pad_token": "<unk>",
126
- "padding_side": "left",
127
  "sp_model_kwargs": {},
128
  "stride": 0,
129
  "tokenizer_class": "LlamaTokenizer",
 
122
  "legacy": false,
123
  "max_length": 512,
124
  "model_max_length": 131072,
125
+ "pad_token": "<|endoftext|>",
126
+ "padding_side": "right",
127
  "sp_model_kwargs": {},
128
  "stride": 0,
129
  "tokenizer_class": "LlamaTokenizer",