Update app.py

app.py

@@ -15,14 +15,28 @@ zip_ref = zipfile.ZipFile(local_zip, 'r')
 zip_ref.extractall('FINAL-EFFICIENTNETV2-B0')
 zip_ref.close()
 
-def deepfakespredict(input_img):
+local_zip = "FINAL-EFFICIENTNETV2-S.zip"
+zip_ref = zipfile.ZipFile(local_zip, 'r')
+zip_ref.extractall('FINAL-EFFICIENTNETV2-S')
+zip_ref.close()
+
+model_b0 = tf.keras.models.load_model("FINAL-EFFICIENTNETV2-B0")
+model_s = tf.keras.models.load_model("FINAL-EFFICIENTNETV2-S")
+
+detector = MTCNN()
 
-    detector = MTCNN()
-    model = tf.keras.models.load_model("FINAL-EFFICIENTNETV2-B0")
-    
-    face = detector.detect_faces(input_img)
 
+def deepfakespredict(input_img, select_model):
+
+    tf.keras.backend.clear_session()
+    
+    if select_model = "EfficientNetV2-B0":
+        model = model_b0
+    elif select_model = "EfficientNetV2-B0":
+        model = model_s
+    
     text =""
+    face = detector.detect_faces(input_img)
 
     if len(face) > 0:
         x, y, width, height = face[0]['box']
@@ -43,24 +57,19 @@ def deepfakespredict(input_img):
             text = "The image is real."
         else:
             text = "The image might be real or fake."
-
-        # if pred[1] >= 0.5:
-        #     text = "The image is fake."
-        # else:
-        #     text = "The image is real."
-
+            
     else:
         text = "Face is not detected in the image."
 
-    return pred, text, input_img
+    return text, input_img, {labels[i]: float(pred[i]) for i in range(2)}
 
 
 title="EfficientNetV2 Deepfakes Image Detector"
 description="This is a demo implementation of EfficientNetV2 Deepfakes Image Detector. To use it, simply upload your image, or click one of the examples to load them."
 examples = []
 gr.Interface(deepfakespredict,
-                     inputs = ["image"],
-                     outputs=["text","text","image"],
+                     inputs = [gr.inputs.Radio(["EfficientNetV2-B0", "EfficientNetV2-S"], label = "Select model:"), "image"],
+                     outputs=["text", gr.outputs.Image(type="pil", label="Detected face"), gr.outputs.Label(num_top_classes=None, type="auto", label="Confidence")],
                      title=title,
                      description=description