app.py

import torch
from TTS.api import TTS
import gradio as gr
import soundfile as sf

# Get device
device = "cuda" if torch.cuda.is_available() else "cpu"

# Load the model from the local path
tts = TTS(model_path="models/xtts_v2/model_file.pth", config_path="models/xtts_v2/config.json").to(device)

# Define available languages
AVAILABLE_LANGUAGES = ["en", "de", "fr", "it", "es", "tr"]

# Function to generate cloned voice
def clone_voice(text, language, speaker_wav):
    if language not in AVAILABLE_LANGUAGES:
        raise ValueError(f"Language '{language}' is not supported. Available languages: {AVAILABLE_LANGUAGES}")
    
    wav = tts.tts(text=text, speaker_wav=speaker_wav, language=language)
    output_file = "output.wav"
    sf.write(output_file, wav, 22050)
    return output_file

# Gradio interface
def gradio_interface(text, language, speaker_wav):
    try:
        output_file = clone_voice(text, language, speaker_wav)
        return output_file
    except Exception as e:
        return str(e)

# Define Gradio inputs and outputs
inputs = [
    gr.Textbox(label="Text to speak", placeholder="Enter text here..."),
    gr.Dropdown(label="Language", choices=AVAILABLE_LANGUAGES, value="en"),
    gr.Audio(label="Reference Voice (Upload or Record)", type="filepath"),
]

outputs = gr.Audio(label="Cloned Voice Output")

# Create Gradio interface
interface = gr.Interface(
    fn=gradio_interface,
    inputs=inputs,
    outputs=outputs,
    title="Voice Cloning with Coqui TTS",
    description="Upload or record a reference voice, enter text, and select a language to generate a cloned voice.",
    live=False,
)

# Launch the interface
interface.launch()