Transformers
pius-code commited on
Commit
bfcfefc
·
verified ·
1 Parent(s): 525d55f

Upload tokenizer

Browse files
Files changed (3) hide show
  1. added_tokens.json +2 -2
  2. tokenizer_config.json +4 -4
  3. vocab.json +52 -33
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 35,
3
- "<s>": 34
4
  }
 
1
  {
2
+ "</s>": 54,
3
+ "<s>": 53
4
  }
tokenizer_config.json CHANGED
@@ -8,7 +8,7 @@
8
  "single_word": false,
9
  "special": true
10
  },
11
- "32": {
12
  "content": "[UNK]",
13
  "lstrip": true,
14
  "normalized": false,
@@ -16,7 +16,7 @@
16
  "single_word": false,
17
  "special": false
18
  },
19
- "33": {
20
  "content": "[PAD]",
21
  "lstrip": true,
22
  "normalized": false,
@@ -24,7 +24,7 @@
24
  "single_word": false,
25
  "special": false
26
  },
27
- "34": {
28
  "content": "<s>",
29
  "lstrip": false,
30
  "normalized": false,
@@ -32,7 +32,7 @@
32
  "single_word": false,
33
  "special": true
34
  },
35
- "35": {
36
  "content": "</s>",
37
  "lstrip": false,
38
  "normalized": false,
 
8
  "single_word": false,
9
  "special": true
10
  },
11
+ "51": {
12
  "content": "[UNK]",
13
  "lstrip": true,
14
  "normalized": false,
 
16
  "single_word": false,
17
  "special": false
18
  },
19
+ "52": {
20
  "content": "[PAD]",
21
  "lstrip": true,
22
  "normalized": false,
 
24
  "single_word": false,
25
  "special": false
26
  },
27
+ "53": {
28
  "content": "<s>",
29
  "lstrip": false,
30
  "normalized": false,
 
32
  "single_word": false,
33
  "special": true
34
  },
35
+ "54": {
36
  "content": "</s>",
37
  "lstrip": false,
38
  "normalized": false,
vocab.json CHANGED
@@ -1,38 +1,57 @@
1
  {
2
  "aka": {
3
- "[PAD]": 33,
4
- "[UNK]": 32,
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
- "p": 16,
21
- "r": 17,
22
- "s": 18,
23
- "t": 19,
24
- "u": 20,
25
- "v": 21,
26
- "w": 22,
27
- "y": 23,
28
- "z": 24,
 
 
 
 
 
 
 
 
29
  "|": 0,
30
- "á": 25,
31
- "ã": 26,
32
- "õ": 27,
33
- "ĩ": 28,
34
- "ũ": 29,
35
- "ɔ": 30,
36
- "ɛ": 31
 
 
 
 
 
 
 
 
 
 
 
37
  }
38
  }
 
1
  {
2
  "aka": {
3
+ ")": 1,
4
+ "0": 2,
5
+ "3": 3,
6
+ "4": 4,
7
+ "5": 5,
8
+ "6": 6,
9
+ "[PAD]": 52,
10
+ "[UNK]": 51,
11
+ "a": 7,
12
+ "b": 8,
13
+ "c": 9,
14
+ "d": 10,
15
+ "e": 11,
16
+ "f": 12,
17
+ "g": 13,
18
+ "h": 14,
19
+ "i": 15,
20
+ "j": 16,
21
+ "k": 17,
22
+ "l": 18,
23
+ "m": 19,
24
+ "n": 20,
25
+ "o": 21,
26
+ "p": 22,
27
+ "q": 23,
28
+ "r": 24,
29
+ "s": 25,
30
+ "t": 26,
31
+ "u": 27,
32
+ "v": 28,
33
+ "w": 29,
34
+ "x": 30,
35
+ "y": 31,
36
+ "z": 32,
37
  "|": 0,
38
+ "·": 33,
39
+ "á": 34,
40
+ "ã": 35,
41
+ "õ": 36,
42
+ "ĩ": 37,
43
+ "ũ": 38,
44
+ "ɔ": 39,
45
+ "ɛ": 40,
46
+ "ʹ": 41,
47
+ "́": 42,
48
+ "—": 43,
49
+ "‘": 44,
50
+ "’": 45,
51
+ "“": 46,
52
+ "•": 47,
53
+ "…": 48,
54
+ "ℇ": 49,
55
+ "ↄ": 50
56
  }
57
  }