jmaczan commited on
Commit
98d81c5
1 Parent(s): 2456367

Upload tokenizer

Browse files
Files changed (3) hide show
  1. added_tokens.json +2 -2
  2. tokenizer_config.json +4 -4
  3. vocab.json +29 -30
added_tokens.json CHANGED
@@ -1,4 +1,4 @@
1
  {
2
- "</s>": 31,
3
- "<s>": 30
4
  }
 
1
  {
2
+ "</s>": 30,
3
+ "<s>": 29
4
  }
tokenizer_config.json CHANGED
@@ -1,6 +1,6 @@
1
  {
2
  "added_tokens_decoder": {
3
- "28": {
4
  "content": "[UNK]",
5
  "lstrip": true,
6
  "normalized": false,
@@ -8,7 +8,7 @@
8
  "single_word": false,
9
  "special": false
10
  },
11
- "29": {
12
  "content": "[PAD]",
13
  "lstrip": true,
14
  "normalized": false,
@@ -16,7 +16,7 @@
16
  "single_word": false,
17
  "special": false
18
  },
19
- "30": {
20
  "content": "<s>",
21
  "lstrip": false,
22
  "normalized": true,
@@ -24,7 +24,7 @@
24
  "single_word": false,
25
  "special": true
26
  },
27
- "31": {
28
  "content": "</s>",
29
  "lstrip": false,
30
  "normalized": true,
 
1
  {
2
  "added_tokens_decoder": {
3
+ "27": {
4
  "content": "[UNK]",
5
  "lstrip": true,
6
  "normalized": false,
 
8
  "single_word": false,
9
  "special": false
10
  },
11
+ "28": {
12
  "content": "[PAD]",
13
  "lstrip": true,
14
  "normalized": false,
 
16
  "single_word": false,
17
  "special": false
18
  },
19
+ "29": {
20
  "content": "<s>",
21
  "lstrip": false,
22
  "normalized": true,
 
24
  "single_word": false,
25
  "special": true
26
  },
27
+ "30": {
28
  "content": "</s>",
29
  "lstrip": false,
30
  "normalized": true,
vocab.json CHANGED
@@ -1,32 +1,31 @@
1
  {
2
- "\u001b": 0,
3
- "[PAD]": 29,
4
- "[UNK]": 28,
5
- "a": 2,
6
- "b": 3,
7
- "c": 4,
8
- "d": 5,
9
- "e": 6,
10
- "f": 7,
11
- "g": 8,
12
- "h": 9,
13
- "i": 10,
14
- "j": 11,
15
- "k": 12,
16
- "l": 13,
17
- "m": 14,
18
- "n": 15,
19
- "o": 16,
20
- "p": 17,
21
- "q": 18,
22
- "r": 19,
23
- "s": 20,
24
- "t": 21,
25
- "u": 22,
26
- "v": 23,
27
- "w": 24,
28
- "x": 25,
29
- "y": 26,
30
- "z": 27,
31
- "|": 1
32
  }
 
1
  {
2
+ "[PAD]": 28,
3
+ "[UNK]": 27,
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
+ "z": 26,
30
+ "|": 0
 
31
  }