Thienpkae commited on
Commit
90416ea
1 Parent(s): f333abc

Upload tokenizer

Browse files
Files changed (3) hide show
  1. added_tokens.json +2 -2
  2. tokenizer_config.json +4 -4
  3. vocab.json +89 -85
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 92,
3
- "<s>": 91
4
  }
 
1
  {
2
+ "</s>": 96,
3
+ "<s>": 95
4
  }
tokenizer_config.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "added_tokens_decoder": {
3
- "89": {
4
  "content": "[UNK]",
5
  "lstrip": true,
6
  "normalized": false,
@@ -8,7 +8,7 @@
8
  "single_word": false,
9
  "special": false
10
  },
11
- "90": {
12
  "content": "[PAD]",
13
  "lstrip": true,
14
  "normalized": false,
@@ -16,7 +16,7 @@
16
  "single_word": false,
17
  "special": false
18
  },
19
- "91": {
20
  "content": "<s>",
21
  "lstrip": false,
22
  "normalized": false,
@@ -24,7 +24,7 @@
24
  "single_word": false,
25
  "special": true
26
  },
27
- "92": {
28
  "content": "</s>",
29
  "lstrip": false,
30
  "normalized": false,
 
1
  {
2
  "added_tokens_decoder": {
3
+ "93": {
4
  "content": "[UNK]",
5
  "lstrip": true,
6
  "normalized": false,
 
8
  "single_word": false,
9
  "special": false
10
  },
11
+ "94": {
12
  "content": "[PAD]",
13
  "lstrip": true,
14
  "normalized": false,
 
16
  "single_word": false,
17
  "special": false
18
  },
19
+ "95": {
20
  "content": "<s>",
21
  "lstrip": false,
22
  "normalized": false,
 
24
  "single_word": false,
25
  "special": true
26
  },
27
+ "96": {
28
  "content": "</s>",
29
  "lstrip": false,
30
  "normalized": false,
vocab.json CHANGED
@@ -1,93 +1,97 @@
1
  {
2
- "[PAD]": 90,
3
- "[UNK]": 89,
4
  "a": 1,
5
  "b": 2,
6
  "c": 3,
7
  "d": 4,
8
  "e": 5,
9
- "g": 6,
10
- "h": 7,
11
- "i": 8,
12
- "j": 9,
13
- "k": 10,
14
- "l": 11,
15
- "m": 12,
16
- "n": 13,
17
- "o": 14,
18
- "p": 15,
19
- "q": 16,
20
- "r": 17,
21
- "s": 18,
22
- "t": 19,
23
- "u": 20,
24
- "v": 21,
25
- "x": 22,
26
- "y": 23,
 
 
27
  "|": 0,
28
- "à": 24,
29
- "á": 25,
30
- "â": 26,
31
- "ã": 27,
32
- "è": 28,
33
- "é": 29,
34
- "ê": 30,
35
- "ì": 31,
36
- "í": 32,
37
- "ò": 33,
38
- "ó": 34,
39
- "ô": 35,
40
- "õ": 36,
41
- "ù": 37,
42
- "ú": 38,
43
- "ý": 39,
44
- "ă": 40,
45
- "đ": 41,
46
- "ĩ": 42,
47
- "ũ": 43,
48
- "ơ": 44,
49
- "ư": 45,
50
- "ạ": 46,
51
- "ả": 47,
52
- "ấ": 48,
53
- "ầ": 49,
54
- "ẩ": 50,
55
- "ẫ": 51,
56
- "ậ": 52,
57
- "ắ": 53,
58
- "ằ": 54,
59
- "ẳ": 55,
60
- "": 56,
61
- "": 57,
62
- "": 58,
63
- "": 59,
64
- "ế": 60,
65
- "": 61,
66
- "": 62,
67
- "": 63,
68
- "": 64,
69
- "": 65,
70
- "": 66,
71
- "": 67,
72
- "": 68,
73
- "": 69,
74
- "": 70,
75
- "": 71,
76
- "": 72,
77
- "": 73,
78
- "": 74,
79
- "": 75,
80
- "": 76,
81
- "": 77,
82
- "": 78,
83
- "": 79,
84
- "": 80,
85
- "": 81,
86
- "": 82,
87
- "": 83,
88
- "": 84,
89
- "": 85,
90
- "": 86,
91
- "": 87,
92
- "": 88
 
 
93
  }
 
1
  {
2
+ "[PAD]": 94,
3
+ "[UNK]": 93,
4
  "a": 1,
5
  "b": 2,
6
  "c": 3,
7
  "d": 4,
8
  "e": 5,
9
+ "f": 6,
10
+ "g": 7,
11
+ "h": 8,
12
+ "i": 9,
13
+ "j": 10,
14
+ "k": 11,
15
+ "l": 12,
16
+ "m": 13,
17
+ "n": 14,
18
+ "o": 15,
19
+ "p": 16,
20
+ "q": 17,
21
+ "r": 18,
22
+ "s": 19,
23
+ "t": 20,
24
+ "u": 21,
25
+ "v": 22,
26
+ "w": 23,
27
+ "x": 24,
28
+ "y": 25,
29
  "|": 0,
30
+ "à": 26,
31
+ "á": 27,
32
+ "â": 28,
33
+ "ã": 29,
34
+ "è": 30,
35
+ "é": 31,
36
+ "ê": 32,
37
+ "ì": 33,
38
+ "í": 34,
39
+ "ò": 35,
40
+ "ó": 36,
41
+ "ô": 37,
42
+ "õ": 38,
43
+ "ù": 39,
44
+ "ú": 40,
45
+ "ý": 41,
46
+ "ă": 42,
47
+ "đ": 43,
48
+ "ĩ": 44,
49
+ "ũ": 45,
50
+ "ơ": 46,
51
+ "ư": 47,
52
+ "ạ": 48,
53
+ "ả": 49,
54
+ "ấ": 50,
55
+ "ầ": 51,
56
+ "ẩ": 52,
57
+ "ẫ": 53,
58
+ "ậ": 54,
59
+ "ắ": 55,
60
+ "ằ": 56,
61
+ "ẳ": 57,
62
+ "": 58,
63
+ "": 59,
64
+ "": 60,
65
+ "": 61,
66
+ "": 62,
67
+ "ế": 63,
68
+ "": 64,
69
+ "": 65,
70
+ "": 66,
71
+ "": 67,
72
+ "": 68,
73
+ "": 69,
74
+ "": 70,
75
+ "": 71,
76
+ "": 72,
77
+ "": 73,
78
+ "": 74,
79
+ "": 75,
80
+ "": 76,
81
+ "": 77,
82
+ "": 78,
83
+ "": 79,
84
+ "": 80,
85
+ "": 81,
86
+ "": 82,
87
+ "": 83,
88
+ "": 84,
89
+ "": 85,
90
+ "": 86,
91
+ "": 87,
92
+ "": 88,
93
+ "": 89,
94
+ "": 90,
95
+ "ỷ": 91,
96
+ "ỹ": 92
97
  }