Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import torch
|
| 3 |
+
from TTS.api import TTS
|
| 4 |
+
import gradio as gr
|
| 5 |
+
from TTS.tts.configs.xtts_config import XttsConfig
|
| 6 |
+
from TTS.tts.models.xtts import Xtts
|
| 7 |
+
from scipy.io.wavfile import write
|
| 8 |
+
import numpy as np
|
| 9 |
+
from transformers import pipeline
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
|
| 13 |
+
def tts_generate(text, speaker_wav="model2.mp3"):
|
| 14 |
+
# Get device
|
| 15 |
+
device = device = "cuda" if torch.cuda.is_available() else "cpu"
|
| 16 |
+
|
| 17 |
+
# Run TTS
|
| 18 |
+
|
| 19 |
+
asr = pipeline("Text-to-Speech", model="coqui/XTTS-v1")
|
| 20 |
+
|
| 21 |
+
# generate speech by cloning a voice using default settings
|
| 22 |
+
tts.tts_to_file(text=text, speaker_wav="voice_models/" + speaker_wav, language="en", file_path="output.wav")
|
| 23 |
+
return "output.wav"
|
| 24 |
+
|
| 25 |
+
def greet(name):
|
| 26 |
+
return "Hello " + name + "!!"
|
| 27 |
+
|
| 28 |
+
iface = gr.Interface(fn=tts_generate,
|
| 29 |
+
inputs=["text", "text"],
|
| 30 |
+
outputs=["audio"],
|
| 31 |
+
examples=[
|
| 32 |
+
["Hello Jhon. Welcome to our group.", "model1.wav"],
|
| 33 |
+
["Hello Jhon. Welcome to our group.", "model2.mp3"]]
|
| 34 |
+
)
|
| 35 |
+
iface.launch(share=True, debug=True)
|
| 36 |
+
|
| 37 |
+
# tts_generate("Hello Jhon. Welcome to our group.", "model2.mp3")
|