Text-to-Speech
English
prince-canuma commited on
Commit
b75823b
·
verified ·
1 Parent(s): 04d6ce5

Upload 2 files

Browse files
Files changed (2) hide show
  1. config.json +1 -150
  2. kokoro-v1_0.safetensors +3 -0
config.json CHANGED
@@ -1,150 +1 @@
1
- {
2
- "istftnet": {
3
- "upsample_kernel_sizes": [20, 12],
4
- "upsample_rates": [10, 6],
5
- "gen_istft_hop_size": 5,
6
- "gen_istft_n_fft": 20,
7
- "resblock_dilation_sizes": [
8
- [1, 3, 5],
9
- [1, 3, 5],
10
- [1, 3, 5]
11
- ],
12
- "resblock_kernel_sizes": [3, 7, 11],
13
- "upsample_initial_channel": 512
14
- },
15
- "dim_in": 64,
16
- "dropout": 0.2,
17
- "hidden_dim": 512,
18
- "max_conv_dim": 512,
19
- "max_dur": 50,
20
- "multispeaker": true,
21
- "n_layer": 3,
22
- "n_mels": 80,
23
- "n_token": 178,
24
- "style_dim": 128,
25
- "text_encoder_kernel_size": 5,
26
- "plbert": {
27
- "hidden_size": 768,
28
- "num_attention_heads": 12,
29
- "intermediate_size": 2048,
30
- "max_position_embeddings": 512,
31
- "num_hidden_layers": 12,
32
- "dropout": 0.1
33
- },
34
- "vocab": {
35
- ";": 1,
36
- ":": 2,
37
- ",": 3,
38
- ".": 4,
39
- "!": 5,
40
- "?": 6,
41
- "—": 9,
42
- "…": 10,
43
- "\"": 11,
44
- "(": 12,
45
- ")": 13,
46
- "“": 14,
47
- "”": 15,
48
- " ": 16,
49
- "\u0303": 17,
50
- "ʣ": 18,
51
- "ʥ": 19,
52
- "ʦ": 20,
53
- "ʨ": 21,
54
- "ᵝ": 22,
55
- "\uAB67": 23,
56
- "A": 24,
57
- "I": 25,
58
- "O": 31,
59
- "Q": 33,
60
- "S": 35,
61
- "T": 36,
62
- "W": 39,
63
- "Y": 41,
64
- "ᵊ": 42,
65
- "a": 43,
66
- "b": 44,
67
- "c": 45,
68
- "d": 46,
69
- "e": 47,
70
- "f": 48,
71
- "h": 50,
72
- "i": 51,
73
- "j": 52,
74
- "k": 53,
75
- "l": 54,
76
- "m": 55,
77
- "n": 56,
78
- "o": 57,
79
- "p": 58,
80
- "q": 59,
81
- "r": 60,
82
- "s": 61,
83
- "t": 62,
84
- "u": 63,
85
- "v": 64,
86
- "w": 65,
87
- "x": 66,
88
- "y": 67,
89
- "z": 68,
90
- "ɑ": 69,
91
- "ɐ": 70,
92
- "ɒ": 71,
93
- "æ": 72,
94
- "β": 75,
95
- "ɔ": 76,
96
- "ɕ": 77,
97
- "ç": 78,
98
- "ɖ": 80,
99
- "ð": 81,
100
- "ʤ": 82,
101
- "ə": 83,
102
- "ɚ": 85,
103
- "ɛ": 86,
104
- "ɜ": 87,
105
- "ɟ": 90,
106
- "ɡ": 92,
107
- "ɥ": 99,
108
- "ɨ": 101,
109
- "ɪ": 102,
110
- "ʝ": 103,
111
- "ɯ": 110,
112
- "ɰ": 111,
113
- "ŋ": 112,
114
- "ɳ": 113,
115
- "ɲ": 114,
116
- "ɴ": 115,
117
- "ø": 116,
118
- "ɸ": 118,
119
- "θ": 119,
120
- "œ": 120,
121
- "ɹ": 123,
122
- "ɾ": 125,
123
- "ɻ": 126,
124
- "ʁ": 128,
125
- "ɽ": 129,
126
- "ʂ": 130,
127
- "ʃ": 131,
128
- "ʈ": 132,
129
- "ʧ": 133,
130
- "ʊ": 135,
131
- "ʋ": 136,
132
- "ʌ": 138,
133
- "ɣ": 139,
134
- "ɤ": 140,
135
- "χ": 142,
136
- "ʎ": 143,
137
- "ʒ": 147,
138
- "ʔ": 148,
139
- "ˈ": 156,
140
- "ˌ": 157,
141
- "ː": 158,
142
- "ʰ": 162,
143
- "ʲ": 164,
144
- "↓": 169,
145
- "→": 171,
146
- "↗": 172,
147
- "↘": 173,
148
- "ᵻ": 177
149
- }
150
- }
 
1
+ {"istftnet": {"upsample_kernel_sizes": [20, 12], "upsample_rates": [10, 6], "gen_istft_hop_size": 5, "gen_istft_n_fft": 20, "resblock_dilation_sizes": [[1, 3, 5], [1, 3, 5], [1, 3, 5]], "resblock_kernel_sizes": [3, 7, 11], "upsample_initial_channel": 512}, "dim_in": 64, "dropout": 0.2, "hidden_dim": 512, "max_conv_dim": 512, "max_dur": 50, "multispeaker": true, "n_layer": 3, "n_mels": 80, "n_token": 178, "style_dim": 128, "text_encoder_kernel_size": 5, "plbert": {"hidden_size": 768, "num_attention_heads": 12, "intermediate_size": 2048, "max_position_embeddings": 512, "num_hidden_layers": 12, "dropout": 0.1}, "vocab": {";": 1, ":": 2, ",": 3, ".": 4, "!": 5, "?": 6, "\u2014": 9, "\u2026": 10, "\"": 11, "(": 12, ")": 13, "\u201c": 14, "\u201d": 15, " ": 16, "\u0303": 17, "\u02a3": 18, "\u02a5": 19, "\u02a6": 20, "\u02a8": 21, "\u1d5d": 22, "\uab67": 23, "A": 24, "I": 25, "O": 31, "Q": 33, "S": 35, "T": 36, "W": 39, "Y": 41, "\u1d4a": 42, "a": 43, "b": 44, "c": 45, "d": 46, "e": 47, "f": 48, "h": 50, "i": 51, "j": 52, "k": 53, "l": 54, "m": 55, "n": 56, "o": 57, "p": 58, "q": 59, "r": 60, "s": 61, "t": 62, "u": 63, "v": 64, "w": 65, "x": 66, "y": 67, "z": 68, "\u0251": 69, "\u0250": 70, "\u0252": 71, "\u00e6": 72, "\u03b2": 75, "\u0254": 76, "\u0255": 77, "\u00e7": 78, "\u0256": 80, "\u00f0": 81, "\u02a4": 82, "\u0259": 83, "\u025a": 85, "\u025b": 86, "\u025c": 87, "\u025f": 90, "\u0261": 92, "\u0265": 99, "\u0268": 101, "\u026a": 102, "\u029d": 103, "\u026f": 110, "\u0270": 111, "\u014b": 112, "\u0273": 113, "\u0272": 114, "\u0274": 115, "\u00f8": 116, "\u0278": 118, "\u03b8": 119, "\u0153": 120, "\u0279": 123, "\u027e": 125, "\u027b": 126, "\u0281": 128, "\u027d": 129, "\u0282": 130, "\u0283": 131, "\u0288": 132, "\u02a7": 133, "\u028a": 135, "\u028b": 136, "\u028c": 138, "\u0263": 139, "\u0264": 140, "\u03c7": 142, "\u028e": 143, "\u0292": 147, "\u0294": 148, "\u02c8": 156, "\u02cc": 157, "\u02d0": 158, "\u02b0": 162, "\u02b2": 164, "\u2193": 169, "\u2192": 171, "\u2197": 172, "\u2198": 173, "\u1d7b": 177}, "quantization": {"group_size": 64, "bits": 8}}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
kokoro-v1_0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fb2b8f22b906f2e70e7dae4c337457784b079fb44142d8f4da93d8a9ace905ed
3
+ size 289324650