[UPDATE]Model

app.py

@@ -7,7 +7,7 @@ import torch.nn as nn
 from PIL import Image
 import requests
 import easyocr
-from transformers import AutoTokenizer
+from transformers import AutoTokenizer, AutoModel
 from torchvision import transforms
 from torchvision import models
 from torchvision.transforms import functional as F
@@ -57,6 +57,18 @@ print("OCR reader initialized.")
 
 # --- Model ---
 
+class TextModelWithClassifier(nn.Module):
+    def __init__(self, base_model):
+        super(TextModelWithClassifier, self).__init__()
+        self.base_model = base_model
+        self.classifier = nn.Linear(base_model.config.hidden_size, 1)
+    
+    def forward(self, input_ids, attention_mask):
+        outputs = self.base_model(input_ids=input_ids, attention_mask=attention_mask)
+        pooled_output = outputs.pooler_output if hasattr(outputs, 'pooler_output') else outputs.last_hidden_state[:, 0]
+        logits = self.classifier(pooled_output)
+        return type('Output', (), {'logits': logits})()
+
 class LateFusionModel(nn.Module):
     def __init__(self, image_model, text_model):
         super(LateFusionModel, self).__init__()
@@ -75,18 +87,31 @@ class LateFusionModel(nn.Module):
 
         return fused_logits, image_logits, text_logits, weights
 
-# Load model
-model_path = "models/best_fusion_model.pt"
+# Load Fusion Model
+# Create model architecture first
+image_model_for_fusion = models.efficientnet_b3(weights=models.EfficientNet_B3_Weights.DEFAULT)
+num_features = image_model_for_fusion.classifier[1].in_features
+image_model_for_fusion.classifier = nn.Linear(num_features, 1)
+
+text_base_model = AutoModel.from_pretrained('indobenchmark/indobert-base-p1')
+text_model = TextModelWithClassifier(text_base_model)
+
+fusion_model = LateFusionModel(image_model_for_fusion, text_model)
+
+# Load state_dict
+model_path = "models/best_mlp_fusion_model_state_dict.pt"
 if os.path.exists(model_path):
-    fusion_model = torch.load(model_path, map_location=device, weights_only=False)
+    fusion_model.load_state_dict(torch.load(model_path, map_location=device))
+    print("Fusion model loaded from local state_dict successfully!")
 else:
-    model_path = hf_hub_download(repo_id="azzandr/gambling-fusion-model", filename="best_fusion_model.pt")
-    fusion_model = torch.load(model_path, map_location=device, weights_only=False)
+    print("Fusion model not found locally. Downloading from Hugging Face Hub...")
+    model_path = hf_hub_download(repo_id="azzandr/gambling-fusion-model", filename="best_mlp_fusion_model_state_dict.pt")
+    fusion_model.load_state_dict(torch.load(model_path, map_location=device))
+    print("Fusion model downloaded and loaded successfully!")
 
-# fusion_model = unwrap_dataparallel(fusion_model)  
 fusion_model.to(device)
 fusion_model.eval()
-print("Fusion model loaded successfully!")
+print("Fusion model ready!")
 
 # Load Image-Only Model
 # Load image model from state_dict