Upload tokenizer
Browse files- special_tokens_map.json +6 -0
- tokenizer.json +115 -0
- tokenizer_config.json +8 -0
- vocab.json +1 -0
special_tokens_map.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": "BOS",
|
3 |
+
"eos_token": "EOS",
|
4 |
+
"pad_token": "PAD",
|
5 |
+
"unk_token": "UNK"
|
6 |
+
}
|
tokenizer.json
ADDED
@@ -0,0 +1,115 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"version": "1.0",
|
3 |
+
"truncation": null,
|
4 |
+
"padding": null,
|
5 |
+
"added_tokens": [
|
6 |
+
{
|
7 |
+
"id": 0,
|
8 |
+
"content": "UNK",
|
9 |
+
"single_word": false,
|
10 |
+
"lstrip": false,
|
11 |
+
"rstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"special": true
|
14 |
+
},
|
15 |
+
{
|
16 |
+
"id": 1,
|
17 |
+
"content": "PAD",
|
18 |
+
"single_word": false,
|
19 |
+
"lstrip": false,
|
20 |
+
"rstrip": false,
|
21 |
+
"normalized": false,
|
22 |
+
"special": true
|
23 |
+
},
|
24 |
+
{
|
25 |
+
"id": 2,
|
26 |
+
"content": "BOS",
|
27 |
+
"single_word": false,
|
28 |
+
"lstrip": false,
|
29 |
+
"rstrip": false,
|
30 |
+
"normalized": false,
|
31 |
+
"special": true
|
32 |
+
},
|
33 |
+
{
|
34 |
+
"id": 3,
|
35 |
+
"content": "EOS",
|
36 |
+
"single_word": false,
|
37 |
+
"lstrip": false,
|
38 |
+
"rstrip": false,
|
39 |
+
"normalized": false,
|
40 |
+
"special": true
|
41 |
+
}
|
42 |
+
],
|
43 |
+
"normalizer": {
|
44 |
+
"type": "Sequence",
|
45 |
+
"normalizers": [
|
46 |
+
{
|
47 |
+
"type": "Replace",
|
48 |
+
"pattern": {
|
49 |
+
"String": "\n"
|
50 |
+
},
|
51 |
+
"content": " UTT_BOUNDARY"
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"type": "Strip",
|
55 |
+
"strip_left": true,
|
56 |
+
"strip_right": true
|
57 |
+
}
|
58 |
+
]
|
59 |
+
},
|
60 |
+
"pre_tokenizer": {
|
61 |
+
"type": "Whitespace"
|
62 |
+
},
|
63 |
+
"post_processor": null,
|
64 |
+
"decoder": null,
|
65 |
+
"model": {
|
66 |
+
"type": "WordLevel",
|
67 |
+
"vocab": {
|
68 |
+
"UNK": 0,
|
69 |
+
"PAD": 1,
|
70 |
+
"BOS": 2,
|
71 |
+
"EOS": 3,
|
72 |
+
"WORD_BOUNDARY": 4,
|
73 |
+
"UTT_BOUNDARY": 5,
|
74 |
+
"oɪ": 6,
|
75 |
+
"a": 7,
|
76 |
+
"ɾ": 8,
|
77 |
+
"k": 9,
|
78 |
+
"t̠ʃ": 10,
|
79 |
+
"i": 11,
|
80 |
+
"s̺": 12,
|
81 |
+
"l": 13,
|
82 |
+
"p": 14,
|
83 |
+
"o": 15,
|
84 |
+
"r": 16,
|
85 |
+
"aɪ": 17,
|
86 |
+
"n": 18,
|
87 |
+
"m": 19,
|
88 |
+
"ð": 20,
|
89 |
+
"e": 21,
|
90 |
+
"ts̻": 22,
|
91 |
+
"β": 23,
|
92 |
+
"s̻": 24,
|
93 |
+
"ʎ": 25,
|
94 |
+
"b": 26,
|
95 |
+
"aʊ": 27,
|
96 |
+
"t": 28,
|
97 |
+
"ɣ": 29,
|
98 |
+
"ɡ": 30,
|
99 |
+
"c": 31,
|
100 |
+
"u": 32,
|
101 |
+
"eɪ": 33,
|
102 |
+
"d": 34,
|
103 |
+
"ts̺": 35,
|
104 |
+
"j": 36,
|
105 |
+
"ɲ": 37,
|
106 |
+
"f": 38,
|
107 |
+
"ʃ": 39,
|
108 |
+
"ɟ": 40,
|
109 |
+
"eʊ": 41,
|
110 |
+
"θ": 42,
|
111 |
+
"x": 43
|
112 |
+
},
|
113 |
+
"unk_token": "UNK"
|
114 |
+
}
|
115 |
+
}
|
tokenizer_config.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": false,
|
3 |
+
"bos_token": "<|endoftext|>",
|
4 |
+
"eos_token": "<|endoftext|>",
|
5 |
+
"model_max_length": 1000000000000000019884624838656,
|
6 |
+
"tokenizer_class": "GPT2Tokenizer",
|
7 |
+
"unk_token": "<|endoftext|>"
|
8 |
+
}
|
vocab.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"UNK":0,"PAD":1,"BOS":2,"EOS":3,"WORD_BOUNDARY":4,"UTT_BOUNDARY":5,"oɪ":6,"a":7,"ɾ":8,"k":9,"t̠ʃ":10,"i":11,"s̺":12,"l":13,"p":14,"o":15,"r":16,"aɪ":17,"n":18,"m":19,"ð":20,"e":21,"ts̻":22,"β":23,"s̻":24,"ʎ":25,"b":26,"aʊ":27,"t":28,"ɣ":29,"ɡ":30,"c":31,"u":32,"eɪ":33,"d":34,"ts̺":35,"j":36,"ɲ":37,"f":38,"ʃ":39,"ɟ":40,"eʊ":41,"θ":42,"x":43}
|