Commit
•
794cf46
1
Parent(s):
54bcdab
Upload tokenizer
Browse files- vocab.json +47 -47
vocab.json
CHANGED
|
@@ -1,51 +1,51 @@
|
|
| 1 |
{
|
| 2 |
"[PAD]": 0,
|
| 3 |
"[UNK]": 48,
|
| 4 |
-
"a":
|
| 5 |
-
"b":
|
| 6 |
-
"d":
|
| 7 |
-
"e":
|
| 8 |
-
"f":
|
| 9 |
-
"h":
|
| 10 |
-
"i":
|
| 11 |
-
"j":
|
| 12 |
-
"k":
|
| 13 |
-
"l":
|
| 14 |
-
"m":
|
| 15 |
-
"n":
|
| 16 |
-
"o":
|
| 17 |
-
"p":
|
| 18 |
-
"r":
|
| 19 |
-
"s":
|
| 20 |
-
"t":
|
| 21 |
-
"u":
|
| 22 |
-
"w":
|
| 23 |
-
"x":
|
| 24 |
-
"y":
|
| 25 |
-
"z":
|
| 26 |
-
"|":
|
| 27 |
-
"ç":
|
| 28 |
-
"ŋ":
|
| 29 |
-
"ɑ":
|
| 30 |
-
"ɒ":
|
| 31 |
-
"ɔ":
|
| 32 |
-
"ə":
|
| 33 |
-
"ɛ":
|
| 34 |
-
"ɜ":
|
| 35 |
-
"ɡ":
|
| 36 |
-
"ɪ":
|
| 37 |
-
"ɸ":
|
| 38 |
-
"ɹ":
|
| 39 |
-
"ɻ":
|
| 40 |
-
"ʂ":
|
| 41 |
-
"ʃ":
|
| 42 |
-
"ʊ":
|
| 43 |
-
"ʋ":
|
| 44 |
-
"ʌ":
|
| 45 |
-
"ʒ":
|
| 46 |
-
"ʔ":
|
| 47 |
-
"ʘ":
|
| 48 |
-
"ʝ":
|
| 49 |
-
"β":
|
| 50 |
-
"θ":
|
| 51 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"[PAD]": 0,
|
| 3 |
"[UNK]": 48,
|
| 4 |
+
"a": 32,
|
| 5 |
+
"b": 35,
|
| 6 |
+
"d": 9,
|
| 7 |
+
"e": 43,
|
| 8 |
+
"f": 46,
|
| 9 |
+
"h": 42,
|
| 10 |
+
"i": 2,
|
| 11 |
+
"j": 25,
|
| 12 |
+
"k": 15,
|
| 13 |
+
"l": 40,
|
| 14 |
+
"m": 30,
|
| 15 |
+
"n": 36,
|
| 16 |
+
"o": 44,
|
| 17 |
+
"p": 6,
|
| 18 |
+
"r": 3,
|
| 19 |
+
"s": 10,
|
| 20 |
+
"t": 18,
|
| 21 |
+
"u": 20,
|
| 22 |
+
"w": 47,
|
| 23 |
+
"x": 5,
|
| 24 |
+
"y": 37,
|
| 25 |
+
"z": 31,
|
| 26 |
+
"|": 34,
|
| 27 |
+
"ç": 11,
|
| 28 |
+
"ŋ": 27,
|
| 29 |
+
"ɑ": 7,
|
| 30 |
+
"ɒ": 29,
|
| 31 |
+
"ɔ": 45,
|
| 32 |
+
"ə": 22,
|
| 33 |
+
"ɛ": 19,
|
| 34 |
+
"ɜ": 41,
|
| 35 |
+
"ɡ": 17,
|
| 36 |
+
"ɪ": 12,
|
| 37 |
+
"ɸ": 16,
|
| 38 |
+
"ɹ": 28,
|
| 39 |
+
"ɻ": 14,
|
| 40 |
+
"ʂ": 33,
|
| 41 |
+
"ʃ": 4,
|
| 42 |
+
"ʊ": 38,
|
| 43 |
+
"ʋ": 1,
|
| 44 |
+
"ʌ": 21,
|
| 45 |
+
"ʒ": 24,
|
| 46 |
+
"ʔ": 8,
|
| 47 |
+
"ʘ": 26,
|
| 48 |
+
"ʝ": 39,
|
| 49 |
+
"β": 13,
|
| 50 |
+
"θ": 23
|
| 51 |
}
|