commit from
Browse files- app.py +31 -22
- logs/ATR/G_74000.pth +0 -3
- requirements.txt +3 -1
app.py
CHANGED
@@ -43,16 +43,23 @@ from scipy.io.wavfile import write
|
|
43 |
import gradio as gr
|
44 |
import scipy.io.wavfile
|
45 |
import numpy as np
|
46 |
-
import
|
47 |
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
|
55 |
hps = utils.get_hparams_from_file("./configs/ATR.json")
|
|
|
56 |
net_g = SynthesizerTrn(
|
57 |
len(symbols),
|
58 |
hps.data.filter_length // 2 + 1,
|
@@ -60,10 +67,21 @@ net_g = SynthesizerTrn(
|
|
60 |
**hps.model)
|
61 |
_ = net_g.eval()
|
62 |
|
63 |
-
_ = utils.load_checkpoint("./
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
64 |
|
65 |
def jtts(text):
|
66 |
-
|
|
|
|
|
|
|
67 |
with torch.no_grad():
|
68 |
x_tst = stn_tst.unsqueeze(0)
|
69 |
x_tst_lengths = torch.LongTensor([stn_tst.size(0)])
|
@@ -71,16 +89,7 @@ def jtts(text):
|
|
71 |
scipy.io.wavfile.write("out.wav", hps.data.sampling_rate, audio)
|
72 |
return "./out.wav"
|
73 |
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
title = "VITS"
|
79 |
-
description = "demo for VITS: Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech. To use it, simply add your text, or click one of the examples to load them. Read more at the links below."
|
80 |
-
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2106.06103'>Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech</a> | <a href='https://github.com/jaywalnut310/vits'>Github Repo</a></p>"
|
81 |
-
|
82 |
-
examples = [
|
83 |
-
["吾輩は猫である。名前はまだない"],
|
84 |
-
["試験勉強頑張ってくださいね"]]
|
85 |
-
|
86 |
-
gr.Interface(jtts, inputs, outputs, title=title, description=description, article=article, examples=examples).launch()
|
|
|
43 |
import gradio as gr
|
44 |
import scipy.io.wavfile
|
45 |
import numpy as np
|
46 |
+
import re
|
47 |
|
48 |
+
jp_match = re.compile(r'.*?[ぁ|あ|ぃ|い|ぅ|う|ぇ|え|ぉ|お|か|が|き|ぎ|く|ぐ|け|げ|こ|ご|さ|ざ|し|じ|す|ず|せ|ぜ|そ|ぞ|た|だ|ち|ぢ|っ|つ|づ|て|で|と|ど|な|に|ぬ|ね|の|は|ば|ぱ|ひ|び|ぴ|ふ|ぶ|ぷ|へ|べ|ぺ|ほ|ぼ|ぽ|ま|み|む|め|も|ゃ|や|ゅ|ゆ|ょ|よ|ら|り|る|れ|ろ|ゎ|わ|ゐ|ゑ|を|ん|ゔ|ゕ|ゖ|ゝ|ゞ|ゟ|゠|ァ|ア|ィ|イ|ゥ|ウ|ェ|エ|ォ|オ|カ|ガ|キ|ギ|ク|グ|ケ|ゲ|コ|ゴ|サ|ザ|シ|ジ|ス|ズ|セ|ゼ|ソ|ゾ|タ|ダ|チ|ヂ|ッ|ツ|ヅ|テ|デ|ト|ド|ナ|ニ|ヌ|ネ|ノ|ハ|バ|パ|ヒ|ビ|ピ|フ|ブ|プ|ヘ|ベ|ペ|ホ|ボ|ポ|マ|ミ|ム|メ|モ|ャ|ヤ|ュ|ユ|ョ|ヨ|ラ|リ|ル|レ|ロ|ヮ|ワ|ヰ|ヱ|ヲ|ン|ヴ|ヵ|ヶ|ヷ|ヸ|ヹ|ヺ]+')
|
49 |
+
|
50 |
+
title = "VITS"
|
51 |
+
description = "demo for VITS: Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech. To use it, simply add your text, or click one of the examples to load them. Read more at the links below."
|
52 |
+
article = "<p style='text-align: center'><a href='https://arxiv.org/abs/2106.06103'>Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech</a> | <a href='https://github.com/jaywalnut310/vits'>Github Repo</a></p>"
|
53 |
+
|
54 |
+
examples = [
|
55 |
+
["原因不明の海面上昇によって、地表の多くが海に沈んだ近未来。"],
|
56 |
+
["幼い頃の事故によって片足を失った少年・斑鳩夏生は、都市での暮らしに見切りを付け、海辺の田舎町へと移り住んだ。"],
|
57 |
+
["身よりのない彼に遺されたのは、海洋地質学者だった祖母の船と潜水艇、そして借金。"],
|
58 |
+
["nanika acltara itsudemo hanashIte kudasai. gakuiNno kotojanaku, shijini kaNsuru kotodemo nanidemo."]
|
59 |
+
]
|
60 |
|
61 |
hps = utils.get_hparams_from_file("./configs/ATR.json")
|
62 |
+
|
63 |
net_g = SynthesizerTrn(
|
64 |
len(symbols),
|
65 |
hps.data.filter_length // 2 + 1,
|
|
|
67 |
**hps.model)
|
68 |
_ = net_g.eval()
|
69 |
|
70 |
+
_ = utils.load_checkpoint("./G_85000.pth", net_g, None)
|
71 |
+
|
72 |
+
|
73 |
+
def get_text(text, hps):
|
74 |
+
text_norm = cleaned_text_to_sequence(text)
|
75 |
+
if hps.data.add_blank:
|
76 |
+
text_norm = commons.intersperse(text_norm, 0)
|
77 |
+
text_norm = torch.LongTensor(text_norm)
|
78 |
+
return text_norm
|
79 |
|
80 |
def jtts(text):
|
81 |
+
if jp_match.match(text):
|
82 |
+
stn_tst = get_text(japanese_phrase_cleaners(text), hps)
|
83 |
+
else:
|
84 |
+
stn_tst = get_text(text, hps)
|
85 |
with torch.no_grad():
|
86 |
x_tst = stn_tst.unsqueeze(0)
|
87 |
x_tst_lengths = torch.LongTensor([stn_tst.size(0)])
|
|
|
89 |
scipy.io.wavfile.write("out.wav", hps.data.sampling_rate, audio)
|
90 |
return "./out.wav"
|
91 |
|
92 |
+
if __name__ == '__main__':
|
93 |
+
inputs = gr.inputs.Textbox(lines=5, label="Input Text")
|
94 |
+
outputs = gr.outputs.Audio(label="Output Audio")
|
95 |
+
gr.Interface(jtts, inputs, outputs, title=title, description=description, article=article, examples=examples).launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
logs/ATR/G_74000.pth
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:03465e5b6a4de27bf15e59b714463f8cf5e188ac018f9225e5380b22645e1c2d
|
3 |
-
size 436558785
|
|
|
|
|
|
|
|
requirements.txt
CHANGED
@@ -8,4 +8,6 @@ tensorboard==2.3.0
|
|
8 |
torch==1.6.0
|
9 |
torchvision==0.7.0
|
10 |
Unidecode==1.1.1
|
11 |
-
|
|
|
|
|
|
8 |
torch==1.6.0
|
9 |
torchvision==0.7.0
|
10 |
Unidecode==1.1.1
|
11 |
+
pyopenjtalk
|
12 |
+
janome
|
13 |
+
regex
|