Upload tokenizer
Browse files- vocab.json +98 -98
vocab.json
CHANGED
@@ -1,105 +1,105 @@
|
|
1 |
{
|
2 |
-
"$":
|
3 |
-
"&":
|
4 |
-
"'":
|
5 |
-
"(":
|
6 |
-
"-":
|
7 |
-
"F":
|
8 |
-
"L":
|
9 |
-
"N":
|
10 |
-
"R":
|
11 |
-
"Y":
|
12 |
-
"[":
|
13 |
"[FR]": 102,
|
14 |
"[FY-NL]": 101,
|
15 |
"[PAD]": 100,
|
16 |
"[UNK]": 99,
|
17 |
-
"]":
|
18 |
-
"a":
|
19 |
-
"b":
|
20 |
-
"c":
|
21 |
-
"d":
|
22 |
-
"e":
|
23 |
-
"f":
|
24 |
-
"g":
|
25 |
-
"h":
|
26 |
-
"i":
|
27 |
-
"j":
|
28 |
-
"k":
|
29 |
-
"l":
|
30 |
-
"m":
|
31 |
-
"n":
|
32 |
-
"o":
|
33 |
-
"p":
|
34 |
-
"q":
|
35 |
-
"r":
|
36 |
-
"s":
|
37 |
-
"t":
|
38 |
-
"u":
|
39 |
-
"v":
|
40 |
-
"w":
|
41 |
-
"x":
|
42 |
-
"y":
|
43 |
-
"z":
|
44 |
-
"|":
|
45 |
-
"~":
|
46 |
-
"«":
|
47 |
-
"´":
|
48 |
-
"»":
|
49 |
-
"à":
|
50 |
-
"á":
|
51 |
-
"â":
|
52 |
-
"ã":
|
53 |
-
"ä":
|
54 |
-
"å":
|
55 |
-
"æ":
|
56 |
-
"ç":
|
57 |
-
"è":
|
58 |
-
"é":
|
59 |
-
"ê":
|
60 |
-
"ë":
|
61 |
-
"í":
|
62 |
-
"î":
|
63 |
-
"ï":
|
64 |
-
"ñ":
|
65 |
-
"ó":
|
66 |
-
"ô":
|
67 |
-
"ö":
|
68 |
-
"ø":
|
69 |
-
"ù":
|
70 |
-
"ú":
|
71 |
-
"û":
|
72 |
-
"ü":
|
73 |
-
"ā":
|
74 |
-
"ă":
|
75 |
-
"ć":
|
76 |
-
"č":
|
77 |
-
"ę":
|
78 |
-
"ĝ":
|
79 |
"ł": 56,
|
80 |
-
"ō":
|
81 |
-
"ő":
|
82 |
-
"œ":
|
83 |
-
"ś":
|
84 |
-
"ş":
|
85 |
-
"š":
|
86 |
-
"ū":
|
87 |
-
"ż":
|
88 |
-
"ž":
|
89 |
-
"ș":
|
90 |
-
"ʿ":
|
91 |
-
"ː":
|
92 |
-
"̀":
|
93 |
-
"́":
|
94 |
-
"̂":
|
95 |
-
"̧":
|
96 |
-
"–":
|
97 |
-
"—":
|
98 |
-
"‘":
|
99 |
-
"’":
|
100 |
-
"…":
|
101 |
-
"ℤ":
|
102 |
-
"保":
|
103 |
-
"宇":
|
104 |
-
"津":
|
105 |
}
|
|
|
1 |
{
|
2 |
+
"$": 98,
|
3 |
+
"&": 51,
|
4 |
+
"'": 31,
|
5 |
+
"(": 3,
|
6 |
+
"-": 82,
|
7 |
+
"F": 45,
|
8 |
+
"L": 11,
|
9 |
+
"N": 20,
|
10 |
+
"R": 22,
|
11 |
+
"Y": 83,
|
12 |
+
"[": 72,
|
13 |
"[FR]": 102,
|
14 |
"[FY-NL]": 101,
|
15 |
"[PAD]": 100,
|
16 |
"[UNK]": 99,
|
17 |
+
"]": 21,
|
18 |
+
"a": 76,
|
19 |
+
"b": 88,
|
20 |
+
"c": 12,
|
21 |
+
"d": 70,
|
22 |
+
"e": 35,
|
23 |
+
"f": 80,
|
24 |
+
"g": 64,
|
25 |
+
"h": 74,
|
26 |
+
"i": 47,
|
27 |
+
"j": 13,
|
28 |
+
"k": 55,
|
29 |
+
"l": 91,
|
30 |
+
"m": 66,
|
31 |
+
"n": 30,
|
32 |
+
"o": 8,
|
33 |
+
"p": 59,
|
34 |
+
"q": 90,
|
35 |
+
"r": 37,
|
36 |
+
"s": 44,
|
37 |
+
"t": 77,
|
38 |
+
"u": 68,
|
39 |
+
"v": 26,
|
40 |
+
"w": 69,
|
41 |
+
"x": 5,
|
42 |
+
"y": 25,
|
43 |
+
"z": 6,
|
44 |
+
"|": 95,
|
45 |
+
"~": 52,
|
46 |
+
"«": 42,
|
47 |
+
"´": 10,
|
48 |
+
"»": 32,
|
49 |
+
"à": 33,
|
50 |
+
"á": 46,
|
51 |
+
"â": 86,
|
52 |
+
"ã": 84,
|
53 |
+
"ä": 9,
|
54 |
+
"å": 39,
|
55 |
+
"æ": 57,
|
56 |
+
"ç": 23,
|
57 |
+
"è": 7,
|
58 |
+
"é": 18,
|
59 |
+
"ê": 15,
|
60 |
+
"ë": 14,
|
61 |
+
"í": 0,
|
62 |
+
"î": 60,
|
63 |
+
"ï": 62,
|
64 |
+
"ñ": 4,
|
65 |
+
"ó": 40,
|
66 |
+
"ô": 65,
|
67 |
+
"ö": 63,
|
68 |
+
"ø": 24,
|
69 |
+
"ù": 71,
|
70 |
+
"ú": 93,
|
71 |
+
"û": 79,
|
72 |
+
"ü": 43,
|
73 |
+
"ā": 94,
|
74 |
+
"ă": 1,
|
75 |
+
"ć": 34,
|
76 |
+
"č": 27,
|
77 |
+
"ę": 92,
|
78 |
+
"ĝ": 48,
|
79 |
"ł": 56,
|
80 |
+
"ō": 81,
|
81 |
+
"ő": 85,
|
82 |
+
"œ": 16,
|
83 |
+
"ś": 49,
|
84 |
+
"ş": 73,
|
85 |
+
"š": 58,
|
86 |
+
"ū": 41,
|
87 |
+
"ż": 50,
|
88 |
+
"ž": 67,
|
89 |
+
"ș": 29,
|
90 |
+
"ʿ": 89,
|
91 |
+
"ː": 36,
|
92 |
+
"̀": 61,
|
93 |
+
"́": 96,
|
94 |
+
"̂": 87,
|
95 |
+
"̧": 28,
|
96 |
+
"–": 19,
|
97 |
+
"—": 17,
|
98 |
+
"‘": 78,
|
99 |
+
"’": 97,
|
100 |
+
"…": 38,
|
101 |
+
"ℤ": 53,
|
102 |
+
"保": 75,
|
103 |
+
"宇": 54,
|
104 |
+
"津": 2
|
105 |
}
|