app.py

import gradio as gr
import time
from transformers import pipeline

def tts_inference(text, model_name):
    model = {"reference": model_name}
    pipe = pipeline("text-to-speech", model=model['reference'])
    print('Processing...')
    t = time.time()
    output = pipe(text)
    t = time.time() - t
    print(f"Took {round(t)} seconds")
    return (output["audio"], output["sampling_rate"])

# List of available TTS models
available_models = [
    "microsoft/speecht5_tts",
    "facebook/mms-tts-eng",
    "suno/bark"
]

gr.Interface(
    fn=tts_inference,
    inputs=[
        gr.Textbox(label="Enter text", placeholder="Type something to convert to speech..."),
        gr.Dropdown(available_models, label="Select Model")
    ],
    outputs=gr.Audio(type="numpy", label="Generated Speech"),
    title="Hugging Face TTS Space",
    description="Enter text and generate speech using Hugging Face's text-to-speech models."
).launch()