Update app.py

app.py

@@ -42,16 +42,35 @@ labels = learn.dls.vocab
 #     pred,pred_idx,probs = learn.predict(img)
 #     return {labels[i]: float(probs[i]) for i in range(len(labels))}
 
-def predict(img):
+
+def predict(img, percent = 0.9):
     img = PILImage.create(img)
     pred, pred_idx, probs = learn.predict(img)
-    results = {labels[i]: float(probs[i]) for i in range(len(labels))}
-    # Adjust results to highlight when 'Survived' meets the 75% threshold
-    if results['Survived'] >= 0.85:
-        results['Survived'] = 1.0  # Indicating high confidence of survival
+    # Get the index of 'Not Survived' and 'Survived' labels
+    idx_not_survived = labels.index('Not Survived')
+    idx_survived = labels.index('Survived')
+    prob_not_survived = probs[idx_not_survived]
+    prob_survived = probs[idx_survived]
+    # Calculate threshold based on desired percent
+    threshold = 1 - percent
+    if prob_not_survived > threshold:
+        return {'Not Survived': float(prob_not_survived)}
     else:
-        results['Survived'] = 0.0  # Indicating it did not meet the threshold
-    return results
+        return {'Survived': float(prob_survived)}
+
+
+
+
+# def predict(img):
+#     img = PILImage.create(img)
+#     pred, pred_idx, probs = learn.predict(img)
+#     results = {labels[i]: float(probs[i]) for i in range(len(labels))}
+#     # Adjust results to highlight when 'Survived' meets the 75% threshold
+#     if results['Survived'] >= 0.85:
+#         results['Survived'] = 1.0  # Indicating high confidence of survival
+#     else:
+#         results['Survived'] = 0.0  # Indicating it did not meet the threshold
+#     return results
 
 
 # def predict(img):