laalays commited on
Commit
8a795f5
1 Parent(s): 566df44

Upload tokenizer

Browse files
Files changed (3) hide show
  1. added_tokens.json +2 -2
  2. tokenizer_config.json +4 -4
  3. vocab.json +26 -36
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 38,
3
- "<s>": 37
4
  }
 
1
  {
2
+ "</s>": 28,
3
+ "<s>": 27
4
  }
tokenizer_config.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "added_tokens_decoder": {
3
- "35": {
4
  "content": "[UNK]",
5
  "lstrip": true,
6
  "normalized": false,
@@ -8,7 +8,7 @@
8
  "single_word": false,
9
  "special": false
10
  },
11
- "36": {
12
  "content": "[PAD]",
13
  "lstrip": true,
14
  "normalized": false,
@@ -16,7 +16,7 @@
16
  "single_word": false,
17
  "special": false
18
  },
19
- "37": {
20
  "content": "<s>",
21
  "lstrip": false,
22
  "normalized": false,
@@ -24,7 +24,7 @@
24
  "single_word": false,
25
  "special": true
26
  },
27
- "38": {
28
  "content": "</s>",
29
  "lstrip": false,
30
  "normalized": false,
 
1
  {
2
  "added_tokens_decoder": {
3
+ "25": {
4
  "content": "[UNK]",
5
  "lstrip": true,
6
  "normalized": false,
 
8
  "single_word": false,
9
  "special": false
10
  },
11
+ "26": {
12
  "content": "[PAD]",
13
  "lstrip": true,
14
  "normalized": false,
 
16
  "single_word": false,
17
  "special": false
18
  },
19
+ "27": {
20
  "content": "<s>",
21
  "lstrip": false,
22
  "normalized": false,
 
24
  "single_word": false,
25
  "special": true
26
  },
27
+ "28": {
28
  "content": "</s>",
29
  "lstrip": false,
30
  "normalized": false,
vocab.json CHANGED
@@ -1,41 +1,31 @@
1
  {
2
  "som": {
3
- "0": 1,
4
- "1": 2,
5
- "2": 3,
6
- "3": 4,
7
- "4": 5,
8
- "5": 6,
9
- "6": 7,
10
- "7": 8,
11
- "8": 9,
12
- "9": 10,
13
- "[PAD]": 36,
14
- "[UNK]": 35,
15
- "a": 11,
16
- "b": 12,
17
- "c": 13,
18
- "d": 14,
19
- "e": 15,
20
- "f": 16,
21
- "g": 17,
22
- "h": 18,
23
- "i": 19,
24
- "j": 20,
25
- "k": 21,
26
- "l": 22,
27
- "m": 23,
28
- "n": 24,
29
- "o": 25,
30
- "q": 26,
31
- "r": 27,
32
- "s": 28,
33
- "t": 29,
34
- "u": 30,
35
- "w": 31,
36
- "x": 32,
37
- "y": 33,
38
  "|": 0,
39
- " ": 34
40
  }
41
  }
 
1
  {
2
  "som": {
3
+ "[PAD]": 26,
4
+ "[UNK]": 25,
5
+ "a": 1,
6
+ "b": 2,
7
+ "c": 3,
8
+ "d": 4,
9
+ "e": 5,
10
+ "f": 6,
11
+ "g": 7,
12
+ "h": 8,
13
+ "i": 9,
14
+ "j": 10,
15
+ "k": 11,
16
+ "l": 12,
17
+ "m": 13,
18
+ "n": 14,
19
+ "o": 15,
20
+ "q": 16,
21
+ "r": 17,
22
+ "s": 18,
23
+ "t": 19,
24
+ "u": 20,
25
+ "w": 21,
26
+ "x": 22,
27
+ "y": 23,
 
 
 
 
 
 
 
 
 
 
28
  "|": 0,
29
+ " ": 24
30
  }
31
  }