Update app.py

app.py

@@ -1,7 +1,37 @@
 import gradio as gr
+from transformers import CLIPModel, AutoTokenizer, RawImage
 
-def greet(name):
-    return "Hello " + name + "!!"
+# Load the CLIP model and tokenizer
+model = CLIPModel.from_pretrained("Xenova/mobileclip_blt")
+tokenizer = AutoTokenizer.from_pretrained("Xenova/mobileclip_blt")
 
-demo = gr.Interface(fn=greet, inputs="text", outputs="text")
-demo.launch()
+# Define the inference function
+def compute_probability(image):
+    # Process the image
+    image = RawImage.read(image)
+    image_inputs = processor(image)
+    image_embeds = vision_model(image_inputs)
+    normalized_image_embeds = image_embeds.normalize().tolist()
+
+    # Compute the probability
+    text_inputs = tokenizer(["cats", "dogs", "birds"], padding="max_length", truncation=True)
+    text_embeds = model(text_inputs)
+    normalized_text_embeds = text_embeds.normalize().tolist()
+
+    probabilities = normalized_image_embeds.map(
+        x => softmax(normalized_text_embeds.map(y => 100 * dot(x, y)))
+    )
+
+    return {"probability": probabilities[0][0]}
+
+# Create the Gradio interface
+iface = gr.Interface(
+    fn=compute_probability,
+    inputs="image",
+    outputs="text",
+    title="CLIP Probability",
+    description="Upload an image and get the probability scores!"
+)
+
+# Launch the interface
+iface.launch()