Upload model.py

model.py

@@ -2,30 +2,6 @@ import torch
 from transformers import BertModel
 
 
-class Ensembler(torch.nn.Module):
-    def __init__(self, specialists):
-        super().__init__()
-        self.specialists = specialists
-
-    def forward(self, input_ids, attention_mask):
-        outputs = []
-
-        for specialist in self.specialists:
-            specialist.eval()
-
-            specialist.to(torch.device(
-                "cuda" if torch.cuda.is_available() else "cpu"))
-
-            outputs.append(specialist(input_ids, attention_mask))
-
-            # Remove the specialist from the GPU
-            specialist.cpu()
-
-        outputs = torch.cat(outputs, dim=1)
-
-        return torch.mean(outputs, dim=1).unsqueeze(1)
-
-
 class LanguageIdentifier(torch.nn.Module):
     def __init__(self):
         super().__init__()
@@ -47,3 +23,32 @@ class LanguageIdentifier(torch.nn.Module):
         outputs = self.linear_layer(outputs)
 
         return outputs
+
+
+class Ensembler(torch.nn.Module):
+    def __init__(self):
+        super().__init__()
+        self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+    def forward(self, input_ids, attention_mask):
+        outputs = []
+
+        for domain in ['politics', 'news', 'law', 'social_media', 'literature', 'web']:
+            specialist = LanguageIdentifier()
+            
+            specialist.load_state_dict(torch.load(f"models/{domain}.pt", map_location=self.device))
+            
+            specialist.eval()
+
+            specialist.to(self.device)
+
+            outputs.append(specialist(input_ids, attention_mask))
+
+            # Remove the specialist from the GPU
+            specialist.cpu()
+            del specialist
+            
+            
+        outputs = torch.cat(outputs, dim=1)
+
+        return torch.mean(outputs, dim=1).unsqueeze(1)