Upload tokenizer
Browse files- added_tokens.json +2 -2
- vocab.json +9 -11
added_tokens.json
CHANGED
|
@@ -1,4 +1,4 @@
|
|
| 1 |
{
|
| 2 |
-
"</s>":
|
| 3 |
-
"<s>":
|
| 4 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"</s>": 31,
|
| 3 |
+
"<s>": 30
|
| 4 |
}
|
vocab.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
{
|
| 2 |
-
"[PAD]":
|
| 3 |
-
"[UNK]":
|
| 4 |
"|": 0,
|
| 5 |
"ا": 1,
|
| 6 |
"ب": 2,
|
|
@@ -22,13 +22,11 @@
|
|
| 22 |
"و": 18,
|
| 23 |
"ِ": 19,
|
| 24 |
"ّ": 20,
|
| 25 |
-
"
|
| 26 |
-
"
|
| 27 |
-
"
|
| 28 |
-
"
|
| 29 |
-
"
|
| 30 |
-
"
|
| 31 |
-
"
|
| 32 |
-
"ی": 28,
|
| 33 |
-
"ﺩ": 29
|
| 34 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"[PAD]": 29,
|
| 3 |
+
"[UNK]": 28,
|
| 4 |
"|": 0,
|
| 5 |
"ا": 1,
|
| 6 |
"ب": 2,
|
|
|
|
| 22 |
"و": 18,
|
| 23 |
"ِ": 19,
|
| 24 |
"ّ": 20,
|
| 25 |
+
"ٔ": 21,
|
| 26 |
+
"پ": 22,
|
| 27 |
+
"چ": 23,
|
| 28 |
+
"ژ": 24,
|
| 29 |
+
"ک": 25,
|
| 30 |
+
"گ": 26,
|
| 31 |
+
"ی": 27
|
|
|
|
|
|
|
| 32 |
}
|