app.py

import streamlit as st
from transformers import BlipForConditionalGeneration, BlipProcessor

# Load the BLIP model and processor
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-large")

def generate_caption(image):
    # Preprocess the image
    pixel_values = processor(images=image, return_tensors="pt").pixel_values
    
    # Generate caption using the BLIP model
    output_ids = model.generate(pixel_values, max_length=50, num_beams=4, early_stopping=True)
    
    # Decode the caption
    caption = processor.decode(output_ids[0], skip_special_tokens=True)
    
    return caption

def main():
    st.title("Image Caption Generator")
    
    # Upload image
    uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])
    
    if uploaded_file is not None:
        # Display the uploaded image
        image = st.image(uploaded_file, caption="Uploaded Image", use_column_width=True)
        
        # Generate caption
        if st.button("Generate Caption"):
            with st.spinner("Generating caption..."):
                caption = generate_caption(uploaded_file.getvalue())
                st.success(f"Caption: {caption}")

if __name__ == "__main__":
    main()