Add first model of the interface and incorporate the CNN model

app.py

@@ -1,12 +1,17 @@
+from model import model
 import gradio as gr
+import numpy as np
 
-def greet(name):
-    return "Hello " + name + "!!"
+def classify_image(image):
+    probabilities = model.predict(image.name)  # Call prediction function
+    labels = ["Category1", "Category2", "Category3", "Category4"]  # Assing labels
+    top_labels = [labels[i] for i in np.argsort(probabilities)[::-1][:4]]
+    top_probs = [round(float(probabilities[i]), 4) for i in np.argsort(probabilities)[::-1][:4]]
 
 demo = gr.Interface(
-    #fn=classify_image,
-    inputs=gr.Image(),
-    outputs=gr.Label(num_top_classes=4),
+    fn=classify_image,
+    inputs=gr.inputs.Image(),
+    outputs=gr.outputs.Label(num_top_classes=4),
     title="Image Classifier",
     description="Upload an image and get the top 4 predicted labels with probabilities.",)
 demo.launch()

model/model.h5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26c4d449632ea072317c16e9d4857e419b67e1b7751f81a89a87c8e75fe9484e
+size 800

model/model.py

@@ -0,0 +1,15 @@
+from keras.models import load_model
+from PIL import Image
+import keras
+import numpy as np
+
+model = load_model('model\model.h5')
+
+def predict(image_path):
+    img = Image.open(image_path)
+    img = img.resize((255, 255))  # Resize to match your model's input size
+    img_array = np.array(img) / 255.0  # Normalize pixel values
+    img_array = np.expand_dims(img_array, axis=0)  # Add batch dimension
+    predictions = model.predict(img_array)
+    return predictions[0]  # Assuming a single output (adjust if needed)
+