app.py

import torch
import gradio as gr
from transformers import pipeline

# Load the ViT+GPT2 image-to-text pipeline with bfloat16 precision
captioner = pipeline(
    "image-to-text",
    model="nlpconnect/vit-gpt2-image-captioning",
    torch_dtype=torch.bfloat16
)

def generate_caption(image):
    """
    Takes a PIL image and returns a generated caption.
    """
    outputs = captioner(image)
    return outputs[0]["generated_text"]

# Build the Gradio interface
with gr.Blocks(theme=gr.themes.Default()) as demo:
    gr.Markdown(
        """
        # 🖼️ Image Caption Generator
        Upload an image to generate a descriptive caption using ViT+GPT2.
        """
    )

    with gr.Row():
        input_image = gr.Image(type="pil", label="Upload Image")
        caption_output = gr.Textbox(label="Generated Caption", lines=2)

    generate_btn = gr.Button("Generate Caption")
    generate_btn.click(fn=generate_caption, inputs=input_image, outputs=caption_output)

    gr.Markdown(
        """
        ---
        Built with 🤗 Transformers (`nlpconnect/vit-gpt2-image-captioning`) and 🚀 Gradio
        """
    )

demo.launch()