Update app.py
Browse files
app.py
CHANGED
@@ -3,6 +3,8 @@ from TTS.api import TTS
|
|
3 |
import gradio as gr
|
4 |
import os
|
5 |
import spaces
|
|
|
|
|
6 |
|
7 |
#inference:
|
8 |
# file_path: "output.wav"
|
@@ -10,42 +12,49 @@ import spaces
|
|
10 |
# language: "en"
|
11 |
|
12 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
13 |
|
14 |
# Agree to Terms of service
|
15 |
# os.environ["COQUI_TOS_AGREED"] = "1"
|
16 |
|
17 |
-
def init_TTS():
|
18 |
# Get device
|
19 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
20 |
# Initialize the TTS model
|
21 |
-
tts = TTS(
|
22 |
return tts
|
23 |
|
24 |
|
25 |
@spaces.GPU
|
26 |
-
def generate_speech(text):
|
27 |
# Generate speech using the provided text, speaker voice, and language
|
28 |
-
file_path = "output.wav"
|
29 |
-
speaker_wav = "content/speaker.wav"
|
30 |
-
language = "en"
|
31 |
tts.tts_to_file(text=text,
|
32 |
-
file_path=file_path,
|
33 |
-
speaker_wav=speaker_wav,
|
34 |
-
language=language)
|
35 |
return file_path
|
36 |
|
37 |
|
38 |
|
39 |
-
def main():
|
40 |
-
|
41 |
-
# call init
|
42 |
-
tts = init_TTS()
|
43 |
|
44 |
# Create the Gradio interface
|
45 |
interface = gr.Interface(
|
46 |
fn=generate_speech,
|
47 |
inputs=[
|
48 |
-
gr.Textbox(label="Enter your text")
|
|
|
49 |
#gr.Textbox(label="Path to target speaker WAV file", value="/content/speaker.wav")
|
50 |
#gr.Dropdown(label="Language", choices=["en"], value="en")
|
51 |
],
|
@@ -59,7 +68,11 @@ def main():
|
|
59 |
return 0
|
60 |
|
61 |
if __name__ == "__main__":
|
62 |
-
|
|
|
|
|
|
|
|
|
63 |
|
64 |
|
65 |
|
|
|
3 |
import gradio as gr
|
4 |
import os
|
5 |
import spaces
|
6 |
+
import yaml
|
7 |
+
|
8 |
|
9 |
#inference:
|
10 |
# file_path: "output.wav"
|
|
|
12 |
# language: "en"
|
13 |
|
14 |
|
15 |
+
def get_config():
|
16 |
+
# get config path
|
17 |
+
config_path = os.environ["CONFIG_PATH"]
|
18 |
+
# Parse the YAML file
|
19 |
+
with open(config_path, 'r') as file:
|
20 |
+
config = yaml.safe_load(file)
|
21 |
+
|
22 |
+
print(config['inference'])
|
23 |
+
print(config_path)
|
24 |
+
return config
|
25 |
+
|
26 |
+
|
27 |
|
28 |
# Agree to Terms of service
|
29 |
# os.environ["COQUI_TOS_AGREED"] = "1"
|
30 |
|
31 |
+
def init_TTS(config):
|
32 |
# Get device
|
33 |
device = "cuda" if torch.cuda.is_available() else "cpu"
|
34 |
# Initialize the TTS model
|
35 |
+
tts = TTS(config['inference']['model']).to(device)
|
36 |
return tts
|
37 |
|
38 |
|
39 |
@spaces.GPU
|
40 |
+
def generate_speech(text, config):
|
41 |
# Generate speech using the provided text, speaker voice, and language
|
|
|
|
|
|
|
42 |
tts.tts_to_file(text=text,
|
43 |
+
file_path=config['inference']['file_path'],
|
44 |
+
speaker_wav=config['inference']['speaker_wav'],
|
45 |
+
language=config['inference']['language'])
|
46 |
return file_path
|
47 |
|
48 |
|
49 |
|
50 |
+
def main(config):
|
|
|
|
|
|
|
51 |
|
52 |
# Create the Gradio interface
|
53 |
interface = gr.Interface(
|
54 |
fn=generate_speech,
|
55 |
inputs=[
|
56 |
+
gr.Textbox(label="Enter your text"),
|
57 |
+
config
|
58 |
#gr.Textbox(label="Path to target speaker WAV file", value="/content/speaker.wav")
|
59 |
#gr.Dropdown(label="Language", choices=["en"], value="en")
|
60 |
],
|
|
|
68 |
return 0
|
69 |
|
70 |
if __name__ == "__main__":
|
71 |
+
# Get config
|
72 |
+
config = get_config()
|
73 |
+
# initialize TTS
|
74 |
+
tts = init_TTS(config)
|
75 |
+
main(config)
|
76 |
|
77 |
|
78 |
|