Create model.py

model.py

@@ -0,0 +1,10 @@
+import torch
+from efficientnet_pytorch import EfficientNet
+
+class_names = ['normal', 'fracture']  # update if more classes
+
+def load_model():
+    model = EfficientNet.from_name("efficientnet-b0", num_classes=len(class_names))
+    model.load_state_dict(torch.load("efnet_fracture_classifier.pth", map_location="cpu"))
+    model.eval()
+    return model