Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import sys
|
| 3 |
+
import subprocess
|
| 4 |
+
|
| 5 |
+
# 1. Clone the repo
|
| 6 |
+
if not os.path.exists("LuxTTS"):
|
| 7 |
+
subprocess.run(["git", "clone", "https://github.com/ysharma3501/LuxTTS.git"])
|
| 8 |
+
|
| 9 |
+
# 2. Install requirements from the cloned folder
|
| 10 |
+
# We use sys.executable to ensure it installs to the Space's current environment
|
| 11 |
+
subprocess.run([sys.executable, "-m", "pip", "install", "-r", "LuxTTS/requirements.txt"])
|
| 12 |
+
|
| 13 |
+
# 3. Add the folder to the Python path
|
| 14 |
+
# This allows 'from zipvoice.luxtts import LuxTTS' to work
|
| 15 |
+
sys.path.append(os.path.abspath("LuxTTS"))
|
| 16 |
+
|
| 17 |
+
import gradio as gr
|
| 18 |
+
import torch
|
| 19 |
+
from zipvoice.luxtts import LuxTTS
|
| 20 |
+
|
| 21 |
+
# Standard setup continues...
|
| 22 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"
|
| 23 |
+
lux_tts = LuxTTS('YatharthS/LuxTTS', device=device, threads=2)
|
| 24 |
+
|
| 25 |
+
def infer(text, audio_prompt, rms, t_shift, num_steps, speed, return_smooth):
|
| 26 |
+
if audio_prompt is None or not text:
|
| 27 |
+
return None
|
| 28 |
+
|
| 29 |
+
encoded_prompt = lux_tts.encode_prompt(audio_prompt, rms=rms)
|
| 30 |
+
final_wav = lux_tts.generate_speech(
|
| 31 |
+
text,
|
| 32 |
+
encoded_prompt,
|
| 33 |
+
num_steps=int(num_steps),
|
| 34 |
+
t_shift=t_shift,
|
| 35 |
+
speed=speed,
|
| 36 |
+
return_smooth=return_smooth
|
| 37 |
+
)
|
| 38 |
+
return (48000, final_wav)
|
| 39 |
+
|
| 40 |
+
# Gradio UI
|
| 41 |
+
with gr.Blocks() as demo:
|
| 42 |
+
gr.Markdown("# LuxTTS (Automatic Setup)")
|
| 43 |
+
with gr.Row():
|
| 44 |
+
with gr.Column():
|
| 45 |
+
input_text = gr.Textbox(label="Text", value="Hello there!")
|
| 46 |
+
input_audio = gr.Audio(label="Reference Voice", type="filepath")
|
| 47 |
+
with gr.Row():
|
| 48 |
+
rms_val = gr.Number(value=0.01, label="RMS")
|
| 49 |
+
t_shift_val = gr.Number(value=0.9, label="T-Shift")
|
| 50 |
+
steps_val = gr.Slider(1, 10, value=4, step=1, label="Steps")
|
| 51 |
+
btn = gr.Button("Synthesize")
|
| 52 |
+
with gr.Column():
|
| 53 |
+
audio_out = gr.Audio(label="Output")
|
| 54 |
+
|
| 55 |
+
btn.click(infer, [input_text, input_audio, rms_val, t_shift_val, steps_val], audio_out)
|
| 56 |
+
|
| 57 |
+
demo.launch()
|