added some imports back so app runs

app.py

@@ -34,7 +34,11 @@ def header_white_bg(text, fontsize = 40, bold = True):
 
 def diagnose_health(file):
     prediction = predict(file)
-    return prediction
+    predicted_strings = []
+    for p in prediction:
+        predicted_string = f"{p['predicted']}, Probability: {float(p['probability']):.2f}"
+        predicted_strings.append(predicted_string)
+    return predicted_strings
 
 def app():
     add_bg_from_local('assets/background.png')

predictor.py

@@ -1,11 +1,16 @@
+import torch
+import torchvision.models as models
+from PIL import Image
+from torchvision import transforms
+
 def predict(image_file):
 
     #load model with params
     model = models.efficientnet_b0(weights=None)
-    model.load_state_dict(torch.load('best_model.pth', map_location=torch.device('cpu')))
+    model.load_state_dict(torch.load('best_model.pth', map_location=torch.device('cpu')), strict=False)
     device = torch.device('cpu')
 
-    classes = [
+    class_names = [
         "Apple___Apple_scab",
         "Apple___Black_rot",
         "Apple___Cedar_apple_rust",
@@ -57,7 +62,7 @@ def predict(image_file):
                         transforms.Normalize(mean=[0.485, 0.456, 0.406],
                                             std=[0.229, 0.224, 0.225])])
         img_normalized = transform(image).unsqueeze(0)
-        
+
         device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
         img_normalized = img_normalized.to(device)
 
@@ -67,8 +72,8 @@ def predict(image_file):
             probs, indices = torch.topk(torch.softmax(output, dim=1), topk)
             # index = output.data.cpu().numpy().argmax()
         tmp_lst = []
-        # print(indices)
-        # print(probs)
+        print(indices)
+        print(probs)
         for j in range(topk):
             tmp_dct = {}
             label_indx = indices[0][j]