colerobertson
commited on
Commit
•
360152d
1
Parent(s):
69f834d
Upload tokenizer
Browse files- added_tokens.json +3 -2
- tokenizer_config.json +3 -3
- vocab.json +1 -3
added_tokens.json
CHANGED
@@ -1,4 +1,5 @@
|
|
1 |
{
|
2 |
-
"</s>":
|
3 |
-
"<s>":
|
|
|
4 |
}
|
|
|
1 |
{
|
2 |
+
"</s>": 31,
|
3 |
+
"<s>": 30,
|
4 |
+
"[UNK]": 29
|
5 |
}
|
tokenizer_config.json
CHANGED
@@ -8,7 +8,7 @@
|
|
8 |
"single_word": false,
|
9 |
"special": false
|
10 |
},
|
11 |
-
"
|
12 |
"content": "[UNK]",
|
13 |
"lstrip": true,
|
14 |
"normalized": false,
|
@@ -16,7 +16,7 @@
|
|
16 |
"single_word": false,
|
17 |
"special": false
|
18 |
},
|
19 |
-
"
|
20 |
"content": "<s>",
|
21 |
"lstrip": false,
|
22 |
"normalized": false,
|
@@ -24,7 +24,7 @@
|
|
24 |
"single_word": false,
|
25 |
"special": true
|
26 |
},
|
27 |
-
"
|
28 |
"content": "</s>",
|
29 |
"lstrip": false,
|
30 |
"normalized": false,
|
|
|
8 |
"single_word": false,
|
9 |
"special": false
|
10 |
},
|
11 |
+
"29": {
|
12 |
"content": "[UNK]",
|
13 |
"lstrip": true,
|
14 |
"normalized": false,
|
|
|
16 |
"single_word": false,
|
17 |
"special": false
|
18 |
},
|
19 |
+
"30": {
|
20 |
"content": "<s>",
|
21 |
"lstrip": false,
|
22 |
"normalized": false,
|
|
|
24 |
"single_word": false,
|
25 |
"special": true
|
26 |
},
|
27 |
+
"31": {
|
28 |
"content": "</s>",
|
29 |
"lstrip": false,
|
30 |
"normalized": false,
|
vocab.json
CHANGED
@@ -5,9 +5,8 @@
|
|
5 |
"2": 33,
|
6 |
"3": 34,
|
7 |
"5": 5,
|
8 |
-
"9": 8,
|
9 |
"[PAD]": 0,
|
10 |
-
"[UNK]":
|
11 |
"a": 15,
|
12 |
"b": 10,
|
13 |
"c": 35,
|
@@ -29,6 +28,5 @@
|
|
29 |
"u": 18,
|
30 |
"w": 24,
|
31 |
"y": 26,
|
32 |
-
"z": 30,
|
33 |
"£": 1
|
34 |
}
|
|
|
5 |
"2": 33,
|
6 |
"3": 34,
|
7 |
"5": 5,
|
|
|
8 |
"[PAD]": 0,
|
9 |
+
"[UNK]": 29,
|
10 |
"a": 15,
|
11 |
"b": 10,
|
12 |
"c": 35,
|
|
|
28 |
"u": 18,
|
29 |
"w": 24,
|
30 |
"y": 26,
|
|
|
31 |
"£": 1
|
32 |
}
|