Commit
•
908497c
1
Parent(s):
cc5af95
Upload tokenizer
Browse files- tokenizer.json +10 -10
tokenizer.json
CHANGED
|
@@ -10,48 +10,48 @@
|
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
| 12 |
"id": 0,
|
| 13 |
-
"special": true,
|
| 14 |
"content": "[PAD]",
|
| 15 |
"single_word": false,
|
| 16 |
"lstrip": false,
|
| 17 |
"rstrip": false,
|
| 18 |
-
"normalized": false
|
|
|
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"id": 100,
|
| 22 |
-
"special": true,
|
| 23 |
"content": "[UNK]",
|
| 24 |
"single_word": false,
|
| 25 |
"lstrip": false,
|
| 26 |
"rstrip": false,
|
| 27 |
-
"normalized": false
|
|
|
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"id": 101,
|
| 31 |
-
"special": true,
|
| 32 |
"content": "[CLS]",
|
| 33 |
"single_word": false,
|
| 34 |
"lstrip": false,
|
| 35 |
"rstrip": false,
|
| 36 |
-
"normalized": false
|
|
|
|
| 37 |
},
|
| 38 |
{
|
| 39 |
"id": 102,
|
| 40 |
-
"special": true,
|
| 41 |
"content": "[SEP]",
|
| 42 |
"single_word": false,
|
| 43 |
"lstrip": false,
|
| 44 |
"rstrip": false,
|
| 45 |
-
"normalized": false
|
|
|
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"id": 103,
|
| 49 |
-
"special": true,
|
| 50 |
"content": "[MASK]",
|
| 51 |
"single_word": false,
|
| 52 |
"lstrip": false,
|
| 53 |
"rstrip": false,
|
| 54 |
-
"normalized": false
|
|
|
|
| 55 |
}
|
| 56 |
],
|
| 57 |
"normalizer": {
|
|
|
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
| 12 |
"id": 0,
|
|
|
|
| 13 |
"content": "[PAD]",
|
| 14 |
"single_word": false,
|
| 15 |
"lstrip": false,
|
| 16 |
"rstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"special": true
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"id": 100,
|
|
|
|
| 22 |
"content": "[UNK]",
|
| 23 |
"single_word": false,
|
| 24 |
"lstrip": false,
|
| 25 |
"rstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"special": true
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"id": 101,
|
|
|
|
| 31 |
"content": "[CLS]",
|
| 32 |
"single_word": false,
|
| 33 |
"lstrip": false,
|
| 34 |
"rstrip": false,
|
| 35 |
+
"normalized": false,
|
| 36 |
+
"special": true
|
| 37 |
},
|
| 38 |
{
|
| 39 |
"id": 102,
|
|
|
|
| 40 |
"content": "[SEP]",
|
| 41 |
"single_word": false,
|
| 42 |
"lstrip": false,
|
| 43 |
"rstrip": false,
|
| 44 |
+
"normalized": false,
|
| 45 |
+
"special": true
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"id": 103,
|
|
|
|
| 49 |
"content": "[MASK]",
|
| 50 |
"single_word": false,
|
| 51 |
"lstrip": false,
|
| 52 |
"rstrip": false,
|
| 53 |
+
"normalized": false,
|
| 54 |
+
"special": true
|
| 55 |
}
|
| 56 |
],
|
| 57 |
"normalizer": {
|