Merge pull request #3698 from guaneec/hn-activation

Remove activation from final layer of Hypernetworks

modules/hypernetworks/hypernetwork.py

@@ -35,7 +35,8 @@ class HypernetworkModule(torch.nn.Module):
     }
     activation_dict.update({cls_name.lower(): cls_obj for cls_name, cls_obj in inspect.getmembers(torch.nn.modules.activation) if inspect.isclass(cls_obj) and cls_obj.__module__ == 'torch.nn.modules.activation'})
 
-    def __init__(self, dim, state_dict=None, layer_structure=None, activation_func=None, weight_init='Normal', add_layer_norm=False, use_dropout=False):
+    def __init__(self, dim, state_dict=None, layer_structure=None, activation_func=None, weight_init='Normal',
+                 add_layer_norm=False, use_dropout=False, activate_output=False, last_layer_dropout=True):
         super().__init__()
 
         assert layer_structure is not None, "layer_structure must not be None"
@@ -48,8 +49,8 @@ class HypernetworkModule(torch.nn.Module):
             # 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) - 2 and not activate_output):
                 pass
             elif activation_func in self.activation_dict:
                 linears.append(self.activation_dict[activation_func]())
@@ -60,8 +61,8 @@ class HypernetworkModule(torch.nn.Module):
             if add_layer_norm:
                 linears.append(torch.nn.LayerNorm(int(dim * layer_structure[i+1])))
 
-            # Add dropout expect last layer
-            if use_dropout and i < len(layer_structure) - 3:
+            # Add dropout except last layer
+            if use_dropout and (i < len(layer_structure) - 3 or last_layer_dropout and i < len(layer_structure) - 2):
                 linears.append(torch.nn.Dropout(p=0.3))
 
         self.linear = torch.nn.Sequential(*linears)
@@ -75,7 +76,7 @@ class HypernetworkModule(torch.nn.Module):
                     w, b = layer.weight.data, layer.bias.data
                     if weight_init == "Normal" or type(layer) == torch.nn.LayerNorm:
                         normal_(w, mean=0.0, std=0.01)
-                        normal_(b, mean=0.0, std=0.005)
+                        normal_(b, mean=0.0, std=0)
                     elif weight_init == 'XavierUniform':
                         xavier_uniform_(w)
                         zeros_(b)
@@ -127,7 +128,7 @@ class Hypernetwork:
     filename = None
     name = None
 
-    def __init__(self, name=None, enable_sizes=None, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False):
+    def __init__(self, name=None, enable_sizes=None, layer_structure=None, activation_func=None, weight_init=None, add_layer_norm=False, use_dropout=False, activate_output=False, **kwargs):
         self.filename = None
         self.name = name
         self.layers = {}
@@ -139,11 +140,15 @@ class Hypernetwork:
         self.weight_init = weight_init
         self.add_layer_norm = add_layer_norm
         self.use_dropout = use_dropout
+        self.activate_output = activate_output
+        self.last_layer_dropout = kwargs['last_layer_dropout'] if 'last_layer_dropout' in kwargs else True
 
         for size in enable_sizes or []:
             self.layers[size] = (
-                HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.weight_init, self.add_layer_norm, self.use_dropout),
-                HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.weight_init, self.add_layer_norm, self.use_dropout),
+                HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.weight_init,
+                                   self.add_layer_norm, self.use_dropout, self.activate_output, last_layer_dropout=self.last_layer_dropout),
+                HypernetworkModule(size, None, self.layer_structure, self.activation_func, self.weight_init,
+                                   self.add_layer_norm, self.use_dropout, self.activate_output, last_layer_dropout=self.last_layer_dropout),
             )
 
     def weights(self):
@@ -171,7 +176,9 @@ class Hypernetwork:
         state_dict['use_dropout'] = self.use_dropout
         state_dict['sd_checkpoint'] = self.sd_checkpoint
         state_dict['sd_checkpoint_name'] = self.sd_checkpoint_name
-
+        state_dict['activate_output'] = self.activate_output
+        state_dict['last_layer_dropout'] = self.last_layer_dropout
+        
         torch.save(state_dict, filename)
 
     def load(self, filename):
@@ -191,12 +198,17 @@ class Hypernetwork:
         print(f"Layer norm is set to {self.add_layer_norm}")
         self.use_dropout = state_dict.get('use_dropout', False)
         print(f"Dropout usage is set to {self.use_dropout}" )
+        self.activate_output = state_dict.get('activate_output', True)
+        print(f"Activate last layer is set to {self.activate_output}")
+        self.last_layer_dropout = state_dict.get('last_layer_dropout', False)
 
         for size, sd in state_dict.items():
             if type(size) == int:
                 self.layers[size] = (
-                    HypernetworkModule(size, sd[0], self.layer_structure, self.activation_func, self.weight_init, self.add_layer_norm, self.use_dropout),
-                    HypernetworkModule(size, sd[1], self.layer_structure, self.activation_func, self.weight_init, self.add_layer_norm, self.use_dropout),
+                    HypernetworkModule(size, sd[0], self.layer_structure, self.activation_func, self.weight_init,
+                                       self.add_layer_norm, self.use_dropout, self.activate_output, last_layer_dropout=self.last_layer_dropout),
+                    HypernetworkModule(size, sd[1], self.layer_structure, self.activation_func, self.weight_init,
+                                       self.add_layer_norm, self.use_dropout, self.activate_output, last_layer_dropout=self.last_layer_dropout),
                 )
 
         self.name = state_dict.get('name', self.name)

modules/ui.py

@@ -1182,8 +1182,8 @@ def create_ui(wrap_gradio_gpu_call):
                     new_hypernetwork_name = gr.Textbox(label="Name")
                     new_hypernetwork_sizes = gr.CheckboxGroup(label="Modules", value=["768", "320", "640", "1280"], choices=["768", "320", "640", "1280"])
                     new_hypernetwork_layer_structure = gr.Textbox("1, 2, 1", label="Enter hypernetwork layer structure", placeholder="1st and last digit must be 1. ex:'1, 2, 1'")
-                    new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork", choices=modules.hypernetworks.ui.keys)
-                    new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. relu-like - Kaiming, sigmoid-like - Xavier is recommended", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"])
+                    new_hypernetwork_activation_func = gr.Dropdown(value="linear", label="Select activation function of hypernetwork. Recommended : Swish / Linear(none)", choices=modules.hypernetworks.ui.keys)
+                    new_hypernetwork_initialization_option = gr.Dropdown(value = "Normal", label="Select Layer weights initialization. Recommended: Kaiming for relu-like, Xavier for sigmoid-like, Normal otherwise", choices=["Normal", "KaimingUniform", "KaimingNormal", "XavierUniform", "XavierNormal"])
                     new_hypernetwork_add_layer_norm = gr.Checkbox(label="Add layer normalization")
                     new_hypernetwork_use_dropout = gr.Checkbox(label="Use dropout")
                     overwrite_old_hypernetwork = gr.Checkbox(value=False, label="Overwrite Old Hypernetwork")