Upload tokenizer
Browse files- tokenizer_config.json +2 -0
- vocab.json +71 -71
tokenizer_config.json
CHANGED
@@ -2,7 +2,9 @@
|
|
2 |
"bos_token": "<s>",
|
3 |
"do_lower_case": false,
|
4 |
"eos_token": "</s>",
|
|
|
5 |
"pad_token": "[PAD]",
|
|
|
6 |
"replace_word_delimiter_char": " ",
|
7 |
"tokenizer_class": "Wav2Vec2CTCTokenizer",
|
8 |
"unk_token": "[UNK]",
|
|
|
2 |
"bos_token": "<s>",
|
3 |
"do_lower_case": false,
|
4 |
"eos_token": "</s>",
|
5 |
+
"max_length": 512,
|
6 |
"pad_token": "[PAD]",
|
7 |
+
"padding": "max_length",
|
8 |
"replace_word_delimiter_char": " ",
|
9 |
"tokenizer_class": "Wav2Vec2CTCTokenizer",
|
10 |
"unk_token": "[UNK]",
|
vocab.json
CHANGED
@@ -1,75 +1,75 @@
|
|
1 |
{
|
2 |
-
"0":
|
3 |
-
"1":
|
4 |
-
"2":
|
5 |
-
"3":
|
6 |
-
"4":
|
7 |
-
"8":
|
8 |
"[PAD]": 72,
|
9 |
"[UNK]": 71,
|
10 |
-
"a":
|
11 |
-
"c":
|
12 |
-
"d":
|
13 |
-
"e":
|
14 |
-
"f":
|
15 |
-
"i":
|
16 |
-
"j":
|
17 |
-
"m":
|
18 |
-
"n":
|
19 |
-
"o":
|
20 |
-
"p":
|
21 |
-
"r":
|
22 |
-
"s":
|
23 |
-
"t":
|
24 |
-
"u":
|
25 |
-
"v":
|
26 |
-
"|":
|
27 |
-
"γ±":
|
28 |
-
"γ²":
|
29 |
-
"γ΄":
|
30 |
-
"γ΅":
|
31 |
-
"γΆ":
|
32 |
-
"γ·":
|
33 |
-
"γΈ":
|
34 |
-
"γΉ":
|
35 |
-
"γΊ":
|
36 |
-
"γ»":
|
37 |
-
"γΌ":
|
38 |
-
"γΎ":
|
39 |
-
"γ
":
|
40 |
-
"γ
":
|
41 |
-
"γ
":
|
42 |
-
"γ
":
|
43 |
-
"γ
":
|
44 |
-
"γ
":
|
45 |
-
"γ
":
|
46 |
-
"γ
":
|
47 |
-
"γ
":
|
48 |
-
"γ
":
|
49 |
-
"γ
":
|
50 |
-
"γ
":
|
51 |
-
"γ
":
|
52 |
-
"γ
":
|
53 |
-
"γ
":
|
54 |
-
"γ
":
|
55 |
-
"γ
":
|
56 |
-
"γ
":
|
57 |
-
"γ
":
|
58 |
-
"γ
":
|
59 |
-
"γ
":
|
60 |
-
"γ
":
|
61 |
-
"γ
":
|
62 |
-
"γ
":
|
63 |
-
"γ
":
|
64 |
-
"γ
":
|
65 |
-
"γ
":
|
66 |
-
"γ
":
|
67 |
-
"γ
":
|
68 |
-
"γ
":
|
69 |
-
"γ
":
|
70 |
-
"γ
":
|
71 |
-
"γ
":
|
72 |
-
"γ
‘":
|
73 |
-
"γ
’":
|
74 |
-
"γ
£":
|
75 |
}
|
|
|
1 |
{
|
2 |
+
"0": 32,
|
3 |
+
"1": 14,
|
4 |
+
"2": 26,
|
5 |
+
"3": 18,
|
6 |
+
"4": 31,
|
7 |
+
"8": 27,
|
8 |
"[PAD]": 72,
|
9 |
"[UNK]": 71,
|
10 |
+
"a": 25,
|
11 |
+
"c": 67,
|
12 |
+
"d": 6,
|
13 |
+
"e": 46,
|
14 |
+
"f": 36,
|
15 |
+
"i": 56,
|
16 |
+
"j": 41,
|
17 |
+
"m": 38,
|
18 |
+
"n": 3,
|
19 |
+
"o": 66,
|
20 |
+
"p": 1,
|
21 |
+
"r": 7,
|
22 |
+
"s": 70,
|
23 |
+
"t": 4,
|
24 |
+
"u": 13,
|
25 |
+
"v": 45,
|
26 |
+
"|": 34,
|
27 |
+
"γ±": 37,
|
28 |
+
"γ²": 12,
|
29 |
+
"γ΄": 61,
|
30 |
+
"γ΅": 53,
|
31 |
+
"γΆ": 19,
|
32 |
+
"γ·": 42,
|
33 |
+
"γΈ": 59,
|
34 |
+
"γΉ": 40,
|
35 |
+
"γΊ": 55,
|
36 |
+
"γ»": 24,
|
37 |
+
"γΌ": 17,
|
38 |
+
"γΎ": 21,
|
39 |
+
"γ
": 68,
|
40 |
+
"γ
": 23,
|
41 |
+
"γ
": 9,
|
42 |
+
"γ
": 39,
|
43 |
+
"γ
": 44,
|
44 |
+
"γ
": 30,
|
45 |
+
"γ
": 69,
|
46 |
+
"γ
": 65,
|
47 |
+
"γ
": 10,
|
48 |
+
"γ
": 11,
|
49 |
+
"γ
": 58,
|
50 |
+
"γ
": 8,
|
51 |
+
"γ
": 35,
|
52 |
+
"γ
": 48,
|
53 |
+
"γ
": 15,
|
54 |
+
"γ
": 52,
|
55 |
+
"γ
": 50,
|
56 |
+
"γ
": 49,
|
57 |
+
"γ
": 20,
|
58 |
+
"γ
": 57,
|
59 |
+
"γ
": 33,
|
60 |
+
"γ
": 28,
|
61 |
+
"γ
": 29,
|
62 |
+
"γ
": 22,
|
63 |
+
"γ
": 5,
|
64 |
+
"γ
": 64,
|
65 |
+
"γ
": 62,
|
66 |
+
"γ
": 2,
|
67 |
+
"γ
": 47,
|
68 |
+
"γ
": 51,
|
69 |
+
"γ
": 43,
|
70 |
+
"γ
": 16,
|
71 |
+
"γ
": 0,
|
72 |
+
"γ
‘": 54,
|
73 |
+
"γ
’": 60,
|
74 |
+
"γ
£": 63
|
75 |
}
|