app.py

import gradio as gr
from transformers import AutoModelForTextToSpeech, AutoTokenizer
import torch

# Load the model and tokenizer
model = AutoModelForTextToSpeech.from_pretrained("parler-tts/parler_tts")
tokenizer = AutoTokenizer.from_pretrained("parler-tts/parler_tts")

# Define the TTS function
def text_to_speech(text):
    inputs = tokenizer(text, return_tensors="pt")
    speech = model.generate(**inputs)
    
    # Convert to an audio format, e.g., 16kHz waveform
    audio = speech[0].cpu().detach().numpy()
    return 16000, audio  # Sample rate, waveform

# Gradio interface
interface = gr.Interface(
    fn=text_to_speech,
    inputs="text",
    outputs="audio",
    title="Text to Speech",
    description="Convert text to speech using the parler-tts/parler_tts model"
)

interface.launch()