Merge pull request #8 from rasbt/ch14-softmax

ch14-softmax update

ch14/ch14_part1.py

@@ -298,7 +298,7 @@ def conv2d(X, W, p=(0, 0), s=(1, 1)):
 mnist_dataset = torchvision.datasets.MNIST(root=image_path, 
                                            train=True, 
                                            transform=transform, 
-                                           download=False)
+                                           download=True)
 
 mnist_valid_dataset = Subset(mnist_dataset, torch.arange(10000)) 
 mnist_train_dataset = Subset(mnist_dataset, torch.arange(10000, len(mnist_dataset)))
@@ -310,6 +310,8 @@ def conv2d(X, W, p=(0, 0), s=(1, 1)):
 
 
 
+
+
 batch_size = 64
 torch.manual_seed(1)
 train_dl = DataLoader(mnist_train_dataset, batch_size, shuffle=True)
@@ -367,8 +369,6 @@ def conv2d(X, W, p=(0, 0), s=(1, 1)):
 model.add_module('dropout', nn.Dropout(p=0.5)) 
 
 model.add_module('fc2', nn.Linear(1024, 10)) 
-model.add_module('softmax', nn.Softmax(dim=1)) 
-
 
 
 
@@ -430,6 +430,8 @@ def train(model, num_epochs, train_dl, valid_dl):
 
 
 
+
+
 x_arr = np.arange(len(hist[0])) + 1
 
 fig = plt.figure(figsize=(12, 4))
@@ -446,6 +448,7 @@ def train(model, num_epochs, train_dl, valid_dl):
 ax.set_xlabel('Epoch', size=15)
 ax.set_ylabel('Accuracy', size=15)
 
+#plt.savefig('figures/14_13.png')
 plt.show()
 
 
@@ -456,7 +459,6 @@ def train(model, num_epochs, train_dl, valid_dl):
 pred = model(mnist_test_dataset.data.unsqueeze(1) / 255.)
 is_correct = (torch.argmax(pred, dim=1) == mnist_test_dataset.targets).float()
 print(f'Test accuracy: {is_correct.mean():.4f}') 
-
 
 
 
@@ -475,6 +477,8 @@ def train(model, num_epochs, train_dl, valid_dl):
             verticalalignment='center', 
             transform=ax.transAxes)
 
+
+plt.savefig('figures/14_14.png')
 plt.show()