Upload tokenizer
Browse files- added_tokens.json +2 -2
- vocab.json +20 -23
added_tokens.json
CHANGED
|
@@ -1,4 +1,4 @@
|
|
| 1 |
{
|
| 2 |
-
"</s>":
|
| 3 |
-
"<s>":
|
| 4 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"</s>": 49,
|
| 3 |
+
"<s>": 48
|
| 4 |
}
|
vocab.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"'": 1,
|
| 3 |
-
"[PAD]":
|
| 4 |
-
"[UNK]":
|
| 5 |
"a": 2,
|
| 6 |
"b": 3,
|
| 7 |
"c": 4,
|
|
@@ -29,25 +29,22 @@
|
|
| 29 |
"y": 26,
|
| 30 |
"z": 27,
|
| 31 |
"|": 0,
|
| 32 |
-
"
|
| 33 |
-
"
|
| 34 |
-
"
|
| 35 |
-
"
|
| 36 |
-
"
|
| 37 |
-
"
|
| 38 |
-
"
|
| 39 |
-
"
|
| 40 |
-
"
|
| 41 |
-
"
|
| 42 |
-
"
|
| 43 |
-
"
|
| 44 |
-
"
|
| 45 |
-
"
|
| 46 |
-
"
|
| 47 |
-
"
|
| 48 |
-
"
|
| 49 |
-
"
|
| 50 |
-
"ŷ": 46,
|
| 51 |
-
"–": 47,
|
| 52 |
-
"—": 48
|
| 53 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"'": 1,
|
| 3 |
+
"[PAD]": 47,
|
| 4 |
+
"[UNK]": 46,
|
| 5 |
"a": 2,
|
| 6 |
"b": 3,
|
| 7 |
"c": 4,
|
|
|
|
| 29 |
"y": 26,
|
| 30 |
"z": 27,
|
| 31 |
"|": 0,
|
| 32 |
+
"à": 28,
|
| 33 |
+
"á": 29,
|
| 34 |
+
"â": 30,
|
| 35 |
+
"ä": 31,
|
| 36 |
+
"é": 32,
|
| 37 |
+
"ê": 33,
|
| 38 |
+
"ë": 34,
|
| 39 |
+
"í": 35,
|
| 40 |
+
"î": 36,
|
| 41 |
+
"ï": 37,
|
| 42 |
+
"ñ": 38,
|
| 43 |
+
"ò": 39,
|
| 44 |
+
"ô": 40,
|
| 45 |
+
"ö": 41,
|
| 46 |
+
"û": 42,
|
| 47 |
+
"ÿ": 43,
|
| 48 |
+
"ŵ": 44,
|
| 49 |
+
"ŷ": 45
|
|
|
|
|
|
|
|
|
|
| 50 |
}
|