foxxy-hm commited on
Commit
985fc62
1 Parent(s): 6cde268

Upload processor

Browse files
Files changed (2) hide show
  1. special_tokens_map.json +16 -0
  2. vocab.json +108 -108
special_tokens_map.json CHANGED
@@ -1,4 +1,20 @@
1
  {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2
  "bos_token": "<s>",
3
  "eos_token": "</s>",
4
  "pad_token": "[PAD]",
 
1
  {
2
+ "additional_special_tokens": [
3
+ {
4
+ "content": "<s>",
5
+ "lstrip": false,
6
+ "normalized": true,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ {
11
+ "content": "</s>",
12
+ "lstrip": false,
13
+ "normalized": true,
14
+ "rstrip": false,
15
+ "single_word": false
16
+ }
17
+ ],
18
  "bos_token": "<s>",
19
  "eos_token": "</s>",
20
  "pad_token": "[PAD]",
vocab.json CHANGED
@@ -1,112 +1,112 @@
1
  {
2
- "%": 25,
3
- "0": 61,
4
- "1": 62,
5
- "2": 44,
6
- "3": 88,
7
- "4": 14,
8
- "5": 22,
9
- "6": 40,
10
- "7": 34,
11
- "8": 85,
12
- "9": 29,
13
  "[PAD]": 109,
14
  "[UNK]": 108,
15
- "a": 2,
16
- "b": 64,
17
- "c": 55,
18
- "d": 106,
19
- "e": 48,
20
- "f": 46,
21
- "g": 26,
22
- "h": 7,
23
- "i": 70,
24
- "k": 83,
25
- "l": 39,
26
- "m": 1,
27
- "n": 84,
28
- "o": 54,
29
- "p": 91,
30
- "q": 105,
31
- "r": 4,
32
- "s": 73,
33
- "t": 6,
34
- "u": 11,
35
- "v": 99,
36
- "w": 43,
37
- "x": 103,
38
- "y": 98,
39
- "z": 49,
40
- "|": 58,
41
- "à": 96,
42
- "á": 28,
43
- "â": 90,
44
- "ã": 95,
45
- "è": 78,
46
- "é": 36,
47
- "ê": 12,
48
- "ì": 33,
49
- "í": 52,
50
- "ò": 50,
51
- "ó": 16,
52
- "ô": 0,
53
- "õ": 74,
54
- "ù": 41,
55
- "ú": 76,
56
- "ý": 79,
57
- "ă": 37,
58
- "đ": 60,
59
- "ĩ": 77,
60
- "ũ": 42,
61
- "ơ": 18,
62
- "ư": 47,
63
- "̀": 19,
64
- "́": 53,
65
- "̣": 57,
66
- "у": 59,
67
- "ạ": 63,
68
- "ả": 97,
69
- "ấ": 56,
70
- "ầ": 66,
71
- "ẩ": 8,
72
- "ẫ": 21,
73
- "ậ": 92,
74
- "ắ": 10,
75
- "ằ": 89,
76
- "ẳ": 80,
77
- "ẵ": 3,
78
- "ặ": 13,
79
- "ẹ": 9,
80
- "ẻ": 38,
81
- "ẽ": 101,
82
- "ế": 100,
83
- "ề": 93,
84
- "ể": 102,
85
- "ễ": 68,
86
- "ệ": 94,
87
- "ỉ": 20,
88
- "ị": 69,
89
- "ọ": 87,
90
- "ỏ": 5,
91
- "ố": 35,
92
- "ồ": 71,
93
- "ổ": 15,
94
- "ỗ": 65,
95
- "ộ": 24,
96
- "ớ": 23,
97
- "ờ": 51,
98
- "ở": 32,
99
- "ỡ": 107,
100
- "ợ": 82,
101
- "ụ": 67,
102
- "ủ": 27,
103
- "ứ": 104,
104
- "ừ": 31,
105
- "ử": 30,
106
- "ữ": 86,
107
- "ự": 45,
108
- "ỳ": 81,
109
- "ỵ": 17,
110
- "ỷ": 75,
111
- "ỹ": 72
112
  }
 
1
  {
2
+ "%": 4,
3
+ "0": 29,
4
+ "1": 97,
5
+ "2": 95,
6
+ "3": 73,
7
+ "4": 64,
8
+ "5": 42,
9
+ "6": 79,
10
+ "7": 99,
11
+ "8": 107,
12
+ "9": 10,
13
  "[PAD]": 109,
14
  "[UNK]": 108,
15
+ "a": 54,
16
+ "b": 21,
17
+ "c": 33,
18
+ "d": 61,
19
+ "e": 17,
20
+ "f": 66,
21
+ "g": 91,
22
+ "h": 88,
23
+ "i": 46,
24
+ "k": 11,
25
+ "l": 77,
26
+ "m": 65,
27
+ "n": 94,
28
+ "o": 40,
29
+ "p": 78,
30
+ "q": 28,
31
+ "r": 53,
32
+ "s": 56,
33
+ "t": 26,
34
+ "u": 3,
35
+ "v": 49,
36
+ "w": 100,
37
+ "x": 60,
38
+ "y": 57,
39
+ "z": 44,
40
+ "|": 34,
41
+ "à": 24,
42
+ "á": 83,
43
+ "â": 85,
44
+ "ã": 96,
45
+ "è": 68,
46
+ "é": 37,
47
+ "ê": 81,
48
+ "ì": 67,
49
+ "í": 62,
50
+ "ò": 25,
51
+ "ó": 52,
52
+ "ô": 82,
53
+ "õ": 89,
54
+ "ù": 76,
55
+ "ú": 58,
56
+ "ý": 9,
57
+ "ă": 98,
58
+ "đ": 48,
59
+ "ĩ": 47,
60
+ "ũ": 12,
61
+ "ơ": 43,
62
+ "ư": 51,
63
+ "̀": 30,
64
+ "́": 2,
65
+ "̣": 70,
66
+ "у": 7,
67
+ "ạ": 38,
68
+ "ả": 50,
69
+ "ấ": 84,
70
+ "ầ": 55,
71
+ "ẩ": 19,
72
+ "ẫ": 36,
73
+ "ậ": 23,
74
+ "ắ": 8,
75
+ "ằ": 18,
76
+ "ẳ": 71,
77
+ "ẵ": 75,
78
+ "ặ": 35,
79
+ "ẹ": 80,
80
+ "ẻ": 90,
81
+ "ẽ": 87,
82
+ "ế": 102,
83
+ "ề": 20,
84
+ "ể": 32,
85
+ "ễ": 13,
86
+ "ệ": 103,
87
+ "ỉ": 106,
88
+ "ị": 31,
89
+ "ọ": 72,
90
+ "ỏ": 0,
91
+ "ố": 41,
92
+ "ồ": 74,
93
+ "ổ": 104,
94
+ "ỗ": 86,
95
+ "ộ": 39,
96
+ "ớ": 27,
97
+ "ờ": 93,
98
+ "ở": 101,
99
+ "ỡ": 6,
100
+ "ợ": 92,
101
+ "ụ": 5,
102
+ "ủ": 16,
103
+ "ứ": 69,
104
+ "ừ": 14,
105
+ "ử": 15,
106
+ "ữ": 63,
107
+ "ự": 22,
108
+ "ỳ": 105,
109
+ "ỵ": 45,
110
+ "ỷ": 1,
111
+ "ỹ": 59
112
  }