Create app.py

app.py

@@ -0,0 +1,47 @@
+import gradio as gr
+import torch
+import torch.nn.functional as F
+from torchvision import transforms
+from PIL import Image
+from model import load_model
+
+# Load models once (inference)
+models = {v: load_model(v) for v in ["f", "c", "q"]}
+
+# Define 2D image preprocessing
+transform = transforms.Compose([
+    transforms.Resize((448, 448)),
+    transforms.ToTensor(),
+    transforms.Normalize(mean=[0.485, 0.456, 0.406],
+                         std=[0.229, 0.224, 0.225])
+])
+
+def predict(img: Image.Image):
+    """
+    Gradio prediction function for a single 2D image.
+    Returns a dict of model-version to positive-class probability.
+    """
+    # Preprocess image and add batch dimension
+    x = transform(img).unsqueeze(0)
+
+    # Run inference and collect probabilities
+    results = {}
+    with torch.no_grad():
+        for version, model in models.items():
+            output = model(x)
+            # assume binary classification: index 1 is positive class
+            prob = F.softmax(output, dim=1)[0, 1].item()
+            results[version] = prob
+    return results
+
+# Build the Gradio interface
+iface = gr.Interface(
+    fn=predict,
+    inputs=gr.Image(type="pil", label="Brain Slice Image (2D)"),
+    outputs=gr.Label(num_top_classes=3, label="Alzheimer Risk Probabilities"),
+    title="Vbai-DPA 2.2c 2D Alzheimer Risk Classification",
+    description="Yüklediğiniz 2D beyin görüntüsü dilimi üzerinden f, c, q modellerinin pozitif sınıf ihtimallerini gösterir."
+)
+
+if __name__ == "__main__":
+    iface.launch()