nielsr HF staff commited on
Commit
474f7ef
1 Parent(s): ed3255e

Upload processor

Browse files
Files changed (2) hide show
  1. special_tokens_map.json +2 -2
  2. tokenizer_config.json +5 -5
special_tokens_map.json CHANGED
@@ -2,7 +2,7 @@
2
  "eos_token": {
3
  "content": "</s>",
4
  "lstrip": true,
5
- "normalized": false,
6
  "rstrip": true,
7
  "single_word": false
8
  },
@@ -16,7 +16,7 @@
16
  "unk_token": {
17
  "content": "<unk>",
18
  "lstrip": true,
19
- "normalized": false,
20
  "rstrip": true,
21
  "single_word": false
22
  }
 
2
  "eos_token": {
3
  "content": "</s>",
4
  "lstrip": true,
5
+ "normalized": true,
6
  "rstrip": true,
7
  "single_word": false
8
  },
 
16
  "unk_token": {
17
  "content": "<unk>",
18
  "lstrip": true,
19
+ "normalized": true,
20
  "rstrip": true,
21
  "single_word": false
22
  }
tokenizer_config.json CHANGED
@@ -3,7 +3,7 @@
3
  "1": {
4
  "content": "</s>",
5
  "lstrip": true,
6
- "normalized": false,
7
  "rstrip": true,
8
  "single_word": false,
9
  "special": true
@@ -11,19 +11,19 @@
11
  "2": {
12
  "content": "<unk>",
13
  "lstrip": true,
14
- "normalized": false,
15
  "rstrip": true,
16
  "single_word": false,
17
  "special": true
18
  }
19
  },
20
- "model_input_names": [
21
- "input_ids"
22
- ],
23
  "additional_special_tokens": [],
24
  "clean_up_tokenization_spaces": true,
25
  "do_lower_case": true,
26
  "eos_token": "</s>",
 
 
 
27
  "model_max_length": 64,
28
  "pad_token": "</s>",
29
  "processor_class": "SiglipProcessor",
 
3
  "1": {
4
  "content": "</s>",
5
  "lstrip": true,
6
+ "normalized": true,
7
  "rstrip": true,
8
  "single_word": false,
9
  "special": true
 
11
  "2": {
12
  "content": "<unk>",
13
  "lstrip": true,
14
+ "normalized": true,
15
  "rstrip": true,
16
  "single_word": false,
17
  "special": true
18
  }
19
  },
 
 
 
20
  "additional_special_tokens": [],
21
  "clean_up_tokenization_spaces": true,
22
  "do_lower_case": true,
23
  "eos_token": "</s>",
24
+ "model_input_names": [
25
+ "input_ids"
26
+ ],
27
  "model_max_length": 64,
28
  "pad_token": "</s>",
29
  "processor_class": "SiglipProcessor",