lgodwangl commited on
Commit
b1f5cb5
·
verified ·
1 Parent(s): 4cef039

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": "</s>",
18
  "lstrip": false,
19
  "normalized": false,
20
  "rstrip": false,
tokenizer_config.json CHANGED
@@ -34,9 +34,9 @@
34
  "clean_up_tokenization_spaces": false,
35
  "eos_token": "</s>",
36
  "legacy": true,
37
- "max_length": 1536,
38
  "model_max_length": 1000000000000000019884624838656,
39
- "pad_token": "<unk>",
40
  "sp_model_kwargs": {},
41
  "spaces_between_special_tokens": false,
42
  "stride": 0,
 
34
  "clean_up_tokenization_spaces": false,
35
  "eos_token": "</s>",
36
  "legacy": true,
37
+ "max_length": 2048,
38
  "model_max_length": 1000000000000000019884624838656,
39
+ "pad_token": "</s>",
40
  "sp_model_kwargs": {},
41
  "spaces_between_special_tokens": false,
42
  "stride": 0,