Commit
•
96e6e1c
1
Parent(s):
dc5d0ba
Upload tokenizer
Browse files- vocab.json +42 -42
vocab.json
CHANGED
|
@@ -1,46 +1,46 @@
|
|
| 1 |
{
|
| 2 |
-
" ":
|
| 3 |
"[PAD]": 43,
|
| 4 |
"[UNK]": 42,
|
| 5 |
-
"a":
|
| 6 |
-
"b":
|
| 7 |
-
"c":
|
| 8 |
-
"d":
|
| 9 |
-
"e":
|
| 10 |
-
"f":
|
| 11 |
-
"g":
|
| 12 |
-
"h":
|
| 13 |
-
"i":
|
| 14 |
-
"j":
|
| 15 |
-
"k":
|
| 16 |
-
"l":
|
| 17 |
-
"m":
|
| 18 |
-
"n":
|
| 19 |
-
"o":
|
| 20 |
-
"p":
|
| 21 |
-
"q":
|
| 22 |
-
"r":
|
| 23 |
-
"s":
|
| 24 |
-
"t":
|
| 25 |
-
"u":
|
| 26 |
-
"v":
|
| 27 |
-
"w":
|
| 28 |
-
"x":
|
| 29 |
-
"y":
|
| 30 |
-
"z":
|
| 31 |
-
"á":
|
| 32 |
-
"â":
|
| 33 |
-
"é":
|
| 34 |
-
"î":
|
| 35 |
-
"ü":
|
| 36 |
-
"ă":
|
| 37 |
-
"ć":
|
| 38 |
-
"č":
|
| 39 |
-
"đ":
|
| 40 |
-
"ş":
|
| 41 |
-
"š":
|
| 42 |
-
"ţ":
|
| 43 |
-
"ș":
|
| 44 |
-
"ț":
|
| 45 |
-
"”":
|
| 46 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
" ": 25,
|
| 3 |
"[PAD]": 43,
|
| 4 |
"[UNK]": 42,
|
| 5 |
+
"a": 21,
|
| 6 |
+
"b": 37,
|
| 7 |
+
"c": 15,
|
| 8 |
+
"d": 12,
|
| 9 |
+
"e": 32,
|
| 10 |
+
"f": 9,
|
| 11 |
+
"g": 28,
|
| 12 |
+
"h": 7,
|
| 13 |
+
"i": 29,
|
| 14 |
+
"j": 38,
|
| 15 |
+
"k": 40,
|
| 16 |
+
"l": 20,
|
| 17 |
+
"m": 8,
|
| 18 |
+
"n": 0,
|
| 19 |
+
"o": 2,
|
| 20 |
+
"p": 10,
|
| 21 |
+
"q": 17,
|
| 22 |
+
"r": 31,
|
| 23 |
+
"s": 26,
|
| 24 |
+
"t": 23,
|
| 25 |
+
"u": 33,
|
| 26 |
+
"v": 22,
|
| 27 |
+
"w": 39,
|
| 28 |
+
"x": 16,
|
| 29 |
+
"y": 41,
|
| 30 |
+
"z": 27,
|
| 31 |
+
"á": 18,
|
| 32 |
+
"â": 3,
|
| 33 |
+
"é": 34,
|
| 34 |
+
"î": 11,
|
| 35 |
+
"ü": 4,
|
| 36 |
+
"ă": 5,
|
| 37 |
+
"ć": 35,
|
| 38 |
+
"č": 24,
|
| 39 |
+
"đ": 36,
|
| 40 |
+
"ş": 19,
|
| 41 |
+
"š": 1,
|
| 42 |
+
"ţ": 6,
|
| 43 |
+
"ș": 14,
|
| 44 |
+
"ț": 13,
|
| 45 |
+
"”": 30
|
| 46 |
}
|