Update app.py

app.py

@@ -1,20 +1,36 @@
-import streamlit as st
 from transformers import pipeline
 from PIL import Image
+import requests
+from io import BytesIO
 
-pipeline = pipeline(task="image-classification", model="julien-c/hotdog-not-hotdog")
+classifier = pipeline("zero-shot-image-classification", model="google/siglip-base-patch16-224")
 
-st.title("Hot Dog? Or Not?")
+st.title("Image classifier model demo")
 
-file_name = st.file_uploader("Upload a hot dog candidate image")
+file_name = st.file_uploader("Upload an image")
+
+def scan_image(image, label, tolerance = 0.01):
+    predictions = classifier(image, candidate_labels = [label, "other"])
+    dict = {}
+    for prediction in predictions:
+        dict[prediction['label']] = prediction['score']
+    # print(json.dumps(dict, indent = 3))
+    return (dict[label] > (dict['other'] + tolerance), dict)
 
 if file_name is not None:
     col1, col2 = st.columns(2)
 
     image = Image.open(file_name)
     col1.image(image, use_column_width=True)
-    predictions = pipeline(image)
 
+    label = st.text_input("What to look for in the image?", "Cats")
+
+    predictions = scan_image(image, label)
+    
     col2.header("Probabilities")
-    for p in predictions:
-        col2.subheader(f"{ p['label'] }: { round(p['score'] * 100, 1)}%")
+    for key, value in predictions[1]:
+        col2.subheader(f"{ key }: { round(value * 100, 1)}%")
+
+    if predictions[1]:
+        st.header("The object is present in the given image")
+    else: st.header("The object is not found in the given image")