Update app.py

app.py

@@ -5,7 +5,7 @@ import matplotlib.pyplot as plt
 import seaborn as sns
 import gradio as gr
 import os
-
+import joblib
 from model import DualStreamTransformer, ArcMarginProduct
 
 css = """
@@ -38,35 +38,34 @@ css = """
 }
 """
 
-DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-MODEL_PATH = "./best_model_fold_5.pt"
 
+FOLD = 5
 model = DualStreamTransformer(n_feat1=25, n_feat2=12, d_model=32).to(DEVICE)
 metric_fc = ArcMarginProduct(32, 2).to(DEVICE)
+checkpoint = torch.load(f"best_model_fold_{FOLD}.pt", map_location=DEVICE)
+model.load_state_dict(checkpoint['model'])
+metric_fc.load_state_dict(checkpoint['fc'])
 
-if os.path.exists(MODEL_PATH):
-    checkpoint = torch.load(MODEL_PATH, map_location=DEVICE)
-    if isinstance(checkpoint, dict) and 'model' in checkpoint:
-        model.load_state_dict(checkpoint['model'])
-        metric_fc.load_state_dict(checkpoint['fc'])
-    else:
-        model.load_state_dict(checkpoint)
-    model.eval()
-    print("模型載入成功！")
 
+scaler_ccmq = joblib.load(f"scaler_ccmq_fold_{FOLD}.pkl")
+scaler_osdi = joblib.load(f"scaler_osdi_fold_{FOLD}.pkl")
 
 def analyze_and_predict(*all_answers):
-    if any(a is None for a in all_answers):
-        raise gr.Error("請完整填寫所有問卷題目！")
-
+    # 1. 數值映射 (與訓練時的編碼一致)
     ccmq_map = {"總是": 5, "經常": 4, "有時": 3, "很少": 2, "沒有": 1}
     osdi_map = {"總是": 4, "經常": 3, "一半一半": 2, "偶而": 1, "完全不曾": 0}
     
-    x1 = torch.tensor([[ccmq_map[a] for a in all_answers[:25]]], dtype=torch.float32).to(DEVICE)
-    x2 = torch.tensor([[osdi_map[a] for a in all_answers[25:]]], dtype=torch.float32).to(DEVICE)
+    x1_raw = np.array([[ccmq_map[a] for a in all_answers[:25]]])
+    x2_raw = np.array([[osdi_map[a] for a in all_answers[25:]]])
 
+    x1_scaled = scaler_ccmq.transform(x1_raw)
+    x2_scaled = scaler_osdi.transform(x2_raw)
+
+    sx1 = torch.tensor(x1_scaled, dtype=torch.float32).to(DEVICE)
+    sx2 = torch.tensor(x2_scaled, dtype=torch.float32).to(DEVICE)
+    
     with torch.no_grad():
-        feats = model(x1, x2)
+        feats = model(sx1, sx2)
         logits = metric_fc.predict(feats)
         probs = torch.softmax(logits, dim=1)
         pred_idx = torch.argmax(probs, dim=1).item()