changed model to require less compute

app.py

@@ -204,7 +204,7 @@ def create_app():
     '''
     # Used to serve with panel serve in command line
     save_dir = FILE_PATH + '/saved_models'
-    base_name = 'tiny_vgg_less_compute'
+    base_name = 'tiny_vgg'
 
     mod_path = f'{save_dir}/{base_name}_model.pth' # Path to the saved model state dict
     settings_path = f'{save_dir}/{base_name}_settings.yaml' # Path to the saved model kwargs

model_training/args.txt

@@ -1,12 +1,12 @@
 --num-workers
 0
 --num-epochs
-25
+100
 --batch-size
 100
 --learning-rate
 0.001
 --patience
-10
+20
 --min-delta
-0.001
+0.0005

model_training/model.py

@@ -69,6 +69,7 @@ class TinyVGG(nn.Module):
         num_convs (int): Number of consecutive convolutional layers + ReLU activations in each VGG block.
         in_channels (int): Number of channels in the input.
         hidden_channels (int): Number of hidden channels between convolutional layers. 
+        fc_hidden_dim (int): Number of output (hidden) features for the first linear layer of the classifer.
         num_classes (int): Number of class labels.
         
     '''
@@ -77,6 +78,7 @@ class TinyVGG(nn.Module):
                  num_convs: int, 
                  in_channels: int, 
                  hidden_channels: int, 
+                 fc_hidden_dim: int,
                  num_classes: int):
         super().__init__()
         
@@ -90,8 +92,7 @@ class TinyVGG(nn.Module):
         self.vgg_body = nn.Sequential(*self.all_blks)
         self.classifier = nn.Sequential(
             nn.Flatten(),
-            nn.LazyLinear(4096), nn.ReLU(), nn.Dropout(0.5),
-            nn.LazyLinear(2048), nn.ReLU(), nn.Dropout(0.5),
+            nn.LazyLinear(fc_hidden_dim), nn.ReLU(), nn.Dropout(0.5),
             nn.LazyLinear(num_classes)
         )
         

model_training/run_training.py

@@ -17,13 +17,13 @@ parser = argparse.ArgumentParser(fromfile_prefix_chars = '@')
 parser.add_argument('-nw', '--num-workers', help = 'Number of workers for dataloaders.',
                     type = int, default = 0)
 parser.add_argument('-ne', '--num-epochs', help = 'Number of epochs to train model for.', 
-                    type = int, default = 15)
+                    type = int, default = 25)
 parser.add_argument('-bs', '--batch-size', help = 'Size of batches to split training set.',
                     type = int, default = 100)
 parser.add_argument('-lr', '--learning-rate', help = 'Learning rate for the optimizer.', 
                     type = float, default = 0.001)
 parser.add_argument('-p', '--patience', help = 'Number of epochs to wait before early stopping.', 
-                    type = int, default = 5)
+                    type = int, default = 10)
 parser.add_argument('-md', '--min-delta', help = 'Minimum decrease in loss to reset patience.', 
                     type = float, default = 0.001)
 
@@ -62,6 +62,7 @@ if __name__ == '__main__':
         'num_convs': 2,
         'in_channels': 1,
         'hidden_channels': 10,
+        'fc_hidden_dim': 64,
         'num_classes': len(train_dl.dataset.classes)
     }
 
@@ -87,4 +88,4 @@ if __name__ == '__main__':
                            device = utils.DEVICE,
                            save_mod = True,
                            save_dir = save_dir,
-                           mod_name = mod_name)
+                           mod_name = mod_name)

saved_models/tiny_vgg_less_compute_model.pth

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:94a16b55d2a65b58c30bcad6dcee77d7e45e15221577795aa4de97a508fddced
-size 38494248

saved_models/tiny_vgg_less_compute_settings.yaml

@@ -1,13 +0,0 @@
-mod_kwargs:
-  hidden_channels: 6
-  in_channels: 1
-  num_blks: 2
-  num_classes: 10
-  num_convs: 2
-train_kwargs:
-  batch_size: 100
-  learning_rate: 0.001
-  min_delta: 0.0005
-  num_epochs: 50
-  num_workers: 0
-  patience: 10

saved_models/tiny_vgg_model.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:10b1913e0c2c44d5a76624196371ae83381a13a939afe9e9e9146354206997e9
-size 41711994
+oid sha256:883adf921aeb35183d6c770234e241280b916d49be5493aefb32d22d7a85dd22
+size 150450

saved_models/tiny_vgg_settings.yaml

@@ -1,4 +1,5 @@
 mod_kwargs:
+  fc_hidden_dim: 64
   hidden_channels: 10
   in_channels: 1
   num_blks: 2
@@ -7,7 +8,7 @@ mod_kwargs:
 train_kwargs:
   batch_size: 100
   learning_rate: 0.001
-  min_delta: 0.001
-  num_epochs: 25
+  min_delta: 0.0005
+  num_epochs: 50
   num_workers: 0
-  patience: 5
+  patience: 15