Spaces:
Runtime error
Runtime error
Karthik64001
commited on
Commit
•
0196eee
1
Parent(s):
40f51c9
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from transformers import VitsModel, AutoTokenizer
|
3 |
+
import torch
|
4 |
+
import scipy.io.wavfile
|
5 |
+
|
6 |
+
# Initialize TTS models
|
7 |
+
tts = TTS(model_name="tts_models/multilingual/multi-dataset/your_tts", progress_bar=False, gpu=False)
|
8 |
+
zh_tts = TTS(model_name="tts_models/zh-CN/baker/tacotron2-DDC-GST", progress_bar=False, gpu=False)
|
9 |
+
de_tts = TTS(model_name="tts_models/de/thorsten/vits", gpu=False)
|
10 |
+
es_tts = TTS(model_name="tts_models/es/mai/tacotron2-DDC", progress_bar=False, gpu=False)
|
11 |
+
tam_tts_model = VitsModel.from_pretrained("facebook/mms-tts-tam")
|
12 |
+
tam_tokenizer = AutoTokenizer.from_pretrained("facebook/mms-tts-tam")
|
13 |
+
|
14 |
+
def text_to_speech(text: str, speaker_wav, speaker_wav_file, language: str):
|
15 |
+
if speaker_wav_file and not speaker_wav:
|
16 |
+
speaker_wav = speaker_wav_file
|
17 |
+
file_path = "output.wav"
|
18 |
+
if language == "zh-CN":
|
19 |
+
zh_tts.tts_to_file(text, file_path=file_path)
|
20 |
+
elif language == "de":
|
21 |
+
de_tts.tts_to_file(text, file_path=file_path)
|
22 |
+
elif language == "es":
|
23 |
+
es_tts.tts_to_file(text, file_path=file_path)
|
24 |
+
elif language == "tam":
|
25 |
+
inputs = tam_tokenizer(text, return_tensors="pt")
|
26 |
+
with torch.no_grad():
|
27 |
+
output = tam_tts_model(**inputs).waveform
|
28 |
+
scipy.io.wavfile.write(file_path, rate=tam_tts_model.config.sampling_rate, data=output.numpy())
|
29 |
+
else:
|
30 |
+
if speaker_wav is not None:
|
31 |
+
tts.tts_to_file(text, speaker_wav=speaker_wav, language=language, file_path=file_path)
|
32 |
+
else:
|
33 |
+
tts.tts_to_file(text, speaker=tts.speakers[0], language=language, file_path=file_path)
|
34 |
+
return file_path
|
35 |
+
|
36 |
+
title = "Voice-Cloning-Demo"
|
37 |
+
|
38 |
+
def toggle(choice):
|
39 |
+
if choice == "mic":
|
40 |
+
return gr.update(visible=True, value=None), gr.update(visible=False, value=None)
|
41 |
+
else:
|
42 |
+
return gr.update(visible=False, value=None), gr.update(visible=True, value=None)
|
43 |
+
|
44 |
+
def handle_language_change(choice):
|
45 |
+
if choice in ["zh-CN", "de", "es", "tam"]:
|
46 |
+
return gr.update(visible=False), gr.update(visible=False), gr.update(visible(False))
|
47 |
+
else:
|
48 |
+
return gr.update(visible=True), gr.update(visible(True)), gr.update(visible(True))
|
49 |
+
|
50 |
+
warming_text = """Please note that Chinese, German, Spanish, and Tamil are currently not supported for voice cloning."""
|
51 |
+
|
52 |
+
with gr.Blocks() as demo:
|
53 |
+
with gr.Row():
|
54 |
+
with gr.Column():
|
55 |
+
text_input = gr.Textbox(label="Input the text", value="", max_lines=3)
|
56 |
+
lan_input = gr.Radio(label="Language", choices=["en", "fr-fr", "pt-br", "zh-CN", "de", "es", "tam"], value="en")
|
57 |
+
gr.Markdown(warming_text)
|
58 |
+
radio = gr.Radio(["mic", "file"], value="mic",
|
59 |
+
label="How would you like to upload your audio?")
|
60 |
+
audio_input_mic = gr.Audio(label="Voice to clone", source="microphone", type="filepath", visible=True)
|
61 |
+
audio_input_file = gr.Audio(label="Voice to clone", type="filepath", visible=False)
|
62 |
+
|
63 |
+
with gr.Row():
|
64 |
+
with gr.Column():
|
65 |
+
btn_clear = gr.Button("Clear")
|
66 |
+
with gr.Column():
|
67 |
+
btn = gr.Button("Submit", variant="primary")
|
68 |
+
with gr.Column():
|
69 |
+
audio_output = gr.Audio(label="Output")
|
70 |
+
|
71 |
+
btn.click(text_to_speech, inputs=[text_input, audio_input_mic,
|
72 |
+
audio_input_file, lan_input], outputs=audio_output)
|
73 |
+
radio.change(toggle, radio, [audio_input_mic, audio_input_file])
|
74 |
+
lan_input.change(handle_language_change, lan_input, [radio, audio_input_mic, audio_input_file])
|
75 |
+
|
76 |
+
demo.launch(enable_queue=True)
|