Update app.py

app.py

@@ -136,16 +136,34 @@ def evaluate_model_with_report(val_loader):
 
 # Performance metrics
 def generate_performance_metrics():
-    y_pred = model.predict(X_test)
-    accuracy = evaluate_model_with_report(val_loader)
-    report = classification_report(y_true, y_pred, target_names=["Ham", "Spam"])
+    model.eval()  # Set model to evaluation mode
+
+    y_true = []  # True labels
+    y_pred = []  # Predicted labels
+
+    with torch.no_grad():
+        for batch in val_loader:
+            inputs = tokenizer(email, padding=True, truncation=True, max_length=256, return_tensors="pt")
+            inputs = {key: val.to(device) for key, val in inputs.items()}
+
+            outputs = model(**inputs)
+            prediction = torch.argmax(outputs.logits, dim=1).item()
+
+            y_true.append(label)
+            y_pred.append(prediction)
+
+    # Compute accuracy and classification report
+    accuracy = accuracy_score(y_true, y_pred)
+    report = classification_report(y_true, y_pred, output_dict=True)
+
     return {
         "accuracy": f"{accuracy:.2%}",
         "precision": f"{report['1']['precision']:.2%}",
         "recall": f"{report['1']['recall']:.2%}",
-        "f1_score": f"{report['1']['f1-score']:.2%}"
+        "f1_score": f"{report['1']['f1-score']:.2%}",
     }
 
+
 # Gradio Interface
 
 def create_interface():