Upload tokenizer
Browse files- vocab.json +130 -130
vocab.json
CHANGED
@@ -1,137 +1,137 @@
|
|
1 |
{
|
2 |
-
"!":
|
3 |
"\"": 128,
|
4 |
-
"&":
|
5 |
-
"'":
|
6 |
-
"(":
|
7 |
-
")":
|
8 |
-
",":
|
9 |
-
"-":
|
10 |
-
".":
|
11 |
-
"/":
|
12 |
-
":":
|
13 |
-
";":
|
14 |
-
"=":
|
15 |
-
"?":
|
16 |
-
"A":
|
17 |
-
"B":
|
18 |
-
"C":
|
19 |
-
"D":
|
20 |
-
"E":
|
21 |
-
"F":
|
22 |
-
"G":
|
23 |
-
"H":
|
24 |
-
"I":
|
25 |
-
"J":
|
26 |
-
"K":
|
27 |
-
"L":
|
28 |
-
"M":
|
29 |
-
"N":
|
30 |
"O": 86,
|
31 |
-
"P":
|
32 |
-
"Q":
|
33 |
-
"R":
|
34 |
-
"S":
|
35 |
-
"T":
|
36 |
-
"U":
|
37 |
-
"V":
|
38 |
-
"W":
|
39 |
-
"X":
|
40 |
-
"Y":
|
41 |
-
"Z":
|
42 |
-
"[":
|
43 |
"[PAD]": 134,
|
44 |
"[UNK]": 133,
|
45 |
-
"]":
|
46 |
-
"a":
|
47 |
-
"b":
|
48 |
-
"c":
|
49 |
-
"d":
|
50 |
-
"e":
|
51 |
-
"f":
|
52 |
"g": 114,
|
53 |
-
"h":
|
54 |
-
"i":
|
55 |
-
"j":
|
56 |
-
"k":
|
57 |
-
"l":
|
58 |
-
"m":
|
59 |
-
"n":
|
60 |
-
"o":
|
61 |
-
"p":
|
62 |
-
"q":
|
63 |
-
"r":
|
64 |
-
"s":
|
65 |
-
"t":
|
66 |
-
"u":
|
67 |
-
"v":
|
68 |
-
"w":
|
69 |
-
"x":
|
70 |
-
"y":
|
71 |
-
"z":
|
72 |
-
"|":
|
73 |
-
"¡":
|
74 |
-
"§":
|
75 |
-
"«":
|
76 |
-
"´":
|
77 |
-
"»":
|
78 |
-
"Á":
|
79 |
-
"Ã":
|
80 |
-
"È":
|
81 |
-
"É":
|
82 |
-
"Ö":
|
83 |
-
"Ú":
|
84 |
-
"ß":
|
85 |
-
"à":
|
86 |
-
"á":
|
87 |
-
"â":
|
88 |
-
"ã":
|
89 |
-
"ä":
|
90 |
-
"æ":
|
91 |
-
"ç":
|
92 |
-
"è":
|
93 |
-
"é":
|
94 |
-
"ê":
|
95 |
-
"ë":
|
96 |
-
"í":
|
97 |
-
"ï":
|
98 |
-
"ð":
|
99 |
-
"ñ":
|
100 |
-
"ò":
|
101 |
-
"ó":
|
102 |
-
"ô":
|
103 |
-
"ö":
|
104 |
-
"ø":
|
105 |
-
"ú":
|
106 |
-
"û":
|
107 |
-
"ü":
|
108 |
-
"þ":
|
109 |
-
"ā":
|
110 |
-
"ć":
|
111 |
-
"č":
|
112 |
-
"ł":
|
113 |
-
"ň":
|
114 |
-
"ō":
|
115 |
-
"Š":
|
116 |
-
"ū":
|
117 |
-
"π":
|
118 |
-
"В":
|
119 |
-
"е":
|
120 |
-
"з":
|
121 |
-
"и":
|
122 |
-
"й":
|
123 |
-
"к":
|
124 |
-
"л":
|
125 |
-
"н":
|
126 |
-
"ь":
|
127 |
-
"я":
|
128 |
-
"ụ":
|
129 |
-
"–":
|
130 |
-
"—":
|
131 |
-
"‘":
|
132 |
-
"’":
|
133 |
-
"“":
|
134 |
-
"”":
|
135 |
-
"…":
|
136 |
-
"€":
|
137 |
}
|
|
|
1 |
{
|
2 |
+
"!": 5,
|
3 |
"\"": 128,
|
4 |
+
"&": 81,
|
5 |
+
"'": 56,
|
6 |
+
"(": 126,
|
7 |
+
")": 118,
|
8 |
+
",": 85,
|
9 |
+
"-": 132,
|
10 |
+
".": 78,
|
11 |
+
"/": 25,
|
12 |
+
":": 51,
|
13 |
+
";": 68,
|
14 |
+
"=": 44,
|
15 |
+
"?": 75,
|
16 |
+
"A": 20,
|
17 |
+
"B": 24,
|
18 |
+
"C": 39,
|
19 |
+
"D": 16,
|
20 |
+
"E": 4,
|
21 |
+
"F": 72,
|
22 |
+
"G": 29,
|
23 |
+
"H": 55,
|
24 |
+
"I": 87,
|
25 |
+
"J": 90,
|
26 |
+
"K": 73,
|
27 |
+
"L": 119,
|
28 |
+
"M": 131,
|
29 |
+
"N": 122,
|
30 |
"O": 86,
|
31 |
+
"P": 106,
|
32 |
+
"Q": 28,
|
33 |
+
"R": 48,
|
34 |
+
"S": 53,
|
35 |
+
"T": 97,
|
36 |
+
"U": 12,
|
37 |
+
"V": 77,
|
38 |
+
"W": 96,
|
39 |
+
"X": 120,
|
40 |
+
"Y": 99,
|
41 |
+
"Z": 47,
|
42 |
+
"[": 41,
|
43 |
"[PAD]": 134,
|
44 |
"[UNK]": 133,
|
45 |
+
"]": 100,
|
46 |
+
"a": 23,
|
47 |
+
"b": 50,
|
48 |
+
"c": 123,
|
49 |
+
"d": 35,
|
50 |
+
"e": 7,
|
51 |
+
"f": 40,
|
52 |
"g": 114,
|
53 |
+
"h": 74,
|
54 |
+
"i": 93,
|
55 |
+
"j": 33,
|
56 |
+
"k": 57,
|
57 |
+
"l": 89,
|
58 |
+
"m": 0,
|
59 |
+
"n": 64,
|
60 |
+
"o": 46,
|
61 |
+
"p": 70,
|
62 |
+
"q": 107,
|
63 |
+
"r": 66,
|
64 |
+
"s": 62,
|
65 |
+
"t": 52,
|
66 |
+
"u": 69,
|
67 |
+
"v": 111,
|
68 |
+
"w": 124,
|
69 |
+
"x": 1,
|
70 |
+
"y": 115,
|
71 |
+
"z": 45,
|
72 |
+
"|": 34,
|
73 |
+
"¡": 19,
|
74 |
+
"§": 36,
|
75 |
+
"«": 21,
|
76 |
+
"´": 30,
|
77 |
+
"»": 2,
|
78 |
+
"Á": 104,
|
79 |
+
"Ã": 31,
|
80 |
+
"È": 8,
|
81 |
+
"É": 76,
|
82 |
+
"Ö": 127,
|
83 |
+
"Ú": 121,
|
84 |
+
"ß": 17,
|
85 |
+
"à": 49,
|
86 |
+
"á": 26,
|
87 |
+
"â": 109,
|
88 |
+
"ã": 108,
|
89 |
+
"ä": 15,
|
90 |
+
"æ": 58,
|
91 |
+
"ç": 13,
|
92 |
+
"è": 79,
|
93 |
+
"é": 9,
|
94 |
+
"ê": 113,
|
95 |
+
"ë": 83,
|
96 |
+
"í": 38,
|
97 |
+
"ï": 95,
|
98 |
+
"ð": 117,
|
99 |
+
"ñ": 80,
|
100 |
+
"ò": 11,
|
101 |
+
"ó": 103,
|
102 |
+
"ô": 54,
|
103 |
+
"ö": 59,
|
104 |
+
"ø": 37,
|
105 |
+
"ú": 91,
|
106 |
+
"û": 71,
|
107 |
+
"ü": 125,
|
108 |
+
"þ": 63,
|
109 |
+
"ā": 129,
|
110 |
+
"ć": 10,
|
111 |
+
"č": 82,
|
112 |
+
"ł": 27,
|
113 |
+
"ň": 110,
|
114 |
+
"ō": 105,
|
115 |
+
"Š": 43,
|
116 |
+
"ū": 130,
|
117 |
+
"π": 60,
|
118 |
+
"В": 84,
|
119 |
+
"е": 22,
|
120 |
+
"з": 94,
|
121 |
+
"и": 101,
|
122 |
+
"й": 67,
|
123 |
+
"к": 14,
|
124 |
+
"л": 116,
|
125 |
+
"н": 42,
|
126 |
+
"ь": 6,
|
127 |
+
"я": 112,
|
128 |
+
"ụ": 98,
|
129 |
+
"–": 88,
|
130 |
+
"—": 65,
|
131 |
+
"‘": 61,
|
132 |
+
"’": 18,
|
133 |
+
"“": 102,
|
134 |
+
"”": 92,
|
135 |
+
"…": 32,
|
136 |
+
"€": 3
|
137 |
}
|