Remove activation from final layer of HNs

Author: guaneec

modules/hypernetworks/hypernetwork.py

@@ -41,8 +41,8 @@ def __init__(self, dim, state_dict=None, layer_structure=None, activation_func=N
             # Add a fully-connected layer
             linears.append(torch.nn.Linear(int(dim * layer_structure[i]), int(dim * layer_structure[i+1])))
 
-            # Add an activation func
-            if activation_func == "linear" or activation_func is None:
+            # Add an activation func except last layer
+            if activation_func == "linear" or activation_func is None or i >= len(layer_structure) - 3:
                 pass
             elif activation_func in self.activation_dict:
                 linears.append(self.activation_dict[activation_func]())
@@ -53,7 +53,7 @@ def __init__(self, dim, state_dict=None, layer_structure=None, activation_func=N
             if add_layer_norm:
                 linears.append(torch.nn.LayerNorm(int(dim * layer_structure[i+1])))
 
-            # Add dropout expect last layer
+            # Add dropout except last layer
             if use_dropout and i < len(layer_structure) - 3:
                 linears.append(torch.nn.Dropout(p=0.3))