app.py

import streamlit as st
from PIL import Image
from transformers import pipeline

# Load your fine-tuned zero-shot model
classifier = pipeline("zero-shot-classification", model="Balajim57/zero-shot-vitb32")

def predict(image, prompt1, prompt2, prompt3):
    # Perform zero-shot classification on the uploaded image with provided prompts
    results = classifier(image, [prompt1, prompt2, prompt3])
    return results["labels"]

# Streamlit UI components
st.title("Zero-Shot Image Classification")
uploaded_image = st.file_uploader("Upload an image", type=["jpg", "jpeg", "png"])
prompt1 = st.text_input("Prompt 1")
prompt2 = st.text_input("Prompt 2")
prompt3 = st.text_input("Prompt 3")
if uploaded_image:
    image = Image.open(uploaded_image)
    st.image(image, caption="Uploaded Image", use_column_width=True)
    if st.button("Classify"):
        results = predict(image, prompt1, prompt2, prompt3)
        st.write("Classification Results:")
        st.write(results)