skydheere commited on
Commit
5a3af2a
·
verified ·
1 Parent(s): 8a3e7d9

Upload tokenizer

Browse files
Files changed (3) hide show
  1. added_tokens.json +2 -2
  2. tokenizer_config.json +4 -4
  3. vocab.json +48 -46
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 48,
3
- "<s>": 47
4
  }
 
1
  {
2
+ "</s>": 50,
3
+ "<s>": 49
4
  }
tokenizer_config.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "added_tokens_decoder": {
3
- "45": {
4
  "content": "[UNK]",
5
  "lstrip": true,
6
  "normalized": false,
@@ -8,7 +8,7 @@
8
  "single_word": false,
9
  "special": false
10
  },
11
- "46": {
12
  "content": "[PAD]",
13
  "lstrip": true,
14
  "normalized": false,
@@ -16,7 +16,7 @@
16
  "single_word": false,
17
  "special": false
18
  },
19
- "47": {
20
  "content": "<s>",
21
  "lstrip": false,
22
  "normalized": false,
@@ -24,7 +24,7 @@
24
  "single_word": false,
25
  "special": true
26
  },
27
- "48": {
28
  "content": "</s>",
29
  "lstrip": false,
30
  "normalized": false,
 
1
  {
2
  "added_tokens_decoder": {
3
+ "47": {
4
  "content": "[UNK]",
5
  "lstrip": true,
6
  "normalized": false,
 
8
  "single_word": false,
9
  "special": false
10
  },
11
+ "48": {
12
  "content": "[PAD]",
13
  "lstrip": true,
14
  "normalized": false,
 
16
  "single_word": false,
17
  "special": false
18
  },
19
+ "49": {
20
  "content": "<s>",
21
  "lstrip": false,
22
  "normalized": false,
 
24
  "single_word": false,
25
  "special": true
26
  },
27
+ "50": {
28
  "content": "</s>",
29
  "lstrip": false,
30
  "normalized": false,
vocab.json CHANGED
@@ -1,51 +1,53 @@
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]": 46,
14
- "[UNK]": 45,
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
- "p": 26,
31
- "q": 27,
32
- "r": 28,
33
- "s": 29,
34
- "t": 30,
35
- "u": 31,
36
- "v": 32,
37
- "w": 33,
38
- "x": 34,
39
- "y": 35,
40
- "z": 36,
 
 
41
  "|": 0,
42
- "á": 37,
43
- "ç": 38,
44
- "é": 39,
45
- "í": 40,
46
- "ó": 41,
47
- "ö": 42,
48
- "ú": 43,
49
- "ü": 44
50
  }
51
  }
 
1
  {
2
  "som": {
3
+ "'": 1,
4
+ ",": 2,
5
+ "0": 3,
6
+ "1": 4,
7
+ "2": 5,
8
+ "3": 6,
9
+ "4": 7,
10
+ "5": 8,
11
+ "6": 9,
12
+ "7": 10,
13
+ "8": 11,
14
+ "9": 12,
15
+ "[PAD]": 48,
16
+ "[UNK]": 47,
17
+ "a": 13,
18
+ "b": 14,
19
+ "c": 15,
20
+ "d": 16,
21
+ "e": 17,
22
+ "f": 18,
23
+ "g": 19,
24
+ "h": 20,
25
+ "i": 21,
26
+ "j": 22,
27
+ "k": 23,
28
+ "l": 24,
29
+ "m": 25,
30
+ "n": 26,
31
+ "o": 27,
32
+ "p": 28,
33
+ "q": 29,
34
+ "r": 30,
35
+ "s": 31,
36
+ "t": 32,
37
+ "u": 33,
38
+ "v": 34,
39
+ "w": 35,
40
+ "x": 36,
41
+ "y": 37,
42
+ "z": 38,
43
  "|": 0,
44
+ "á": 39,
45
+ "ç": 40,
46
+ "é": 41,
47
+ "í": 42,
48
+ "ó": 43,
49
+ "ö": 44,
50
+ "ú": 45,
51
+ "ü": 46
52
  }
53
  }