app.py

import gradio as gr
import torch
import os
from dalle_mini import DalleBart
from transformers import pipeline
from huggingface_hub import InferenceClient
from huggingface_hub import login

login(token=os.getenv('HF_TOKEN'))

pipe = DalleBart.from_pretrained("dalle-mini/dalle-mini", dtype="auto")

def image_gen(prompt):
    image = pipe(prompt).images[0]

iface = gr.Interface(
    fn=image_gen,
    inputs=gr.Textbox(lines=2,placeholder="Enter your prompt here..."),
    outputs=gr.Image(type="pil"),
    title="Text-to-Image Generator with VLM",
    description="Generate images from text prompts using a Vision-Language Model."
    )