Tag: gradient-vanishing

# nn layers
linear1 = torch.nn.Linear(784, 32, bias=True)
linear2 = torch.nn.Linear(32, 32, bias=True)
linear3 = torch.nn.Linear(32, 10, bias=True)
relu = torch.nn.ReLU()
bn1 = torch.nn.BatchNorm1d(32)
bn2 = torch.nn.BatchNorm1d(32)

nn_linear1 = torch.nn.Linear(784, 32, bias=True)
nn_linear2 = torch.nn.Linear(32, 32, bias=True)
nn_linear3 = torch.nn.Linear(32, 10, bias=True)

# model with Batch normalization
bn_model = torch.nn.Sequential(linear1, bn1, relu,
                            linear2, bn2, relu,
                            linear3).to(device)

# model without Batch normalization
nn_model = torch.nn.Sequential(nn_linear1, relu,
                               nn_linear2, relu,
                               nn_linear3).to(device)

# define cost/loss & optimizer
criterion = torch.nn.CrossEntropyLoss().to(device)    # Softmax is internally computed.
bn_optimizer = torch.optim.Adam(bn_model.parameters(), lr=learning_rate)
nn_optimizer = torch.optim.Adam(nn_model.parameters(), lr=learning_rate)

criterion = torch.nn.CrossEntropyLoss().to(device)
optimizer = torch.optim.Adam(linear.parameters(), lr=learning_rate) # Adam optimizer

total_batch = len(data_loader)
for epoch in range(training_epochs):
    avg_cost = 0

    for X, Y in data_loader:
        # reshape input image into [batch_size by 784]
        # label is not one-hot encoded
        X = X.view(-1, 28 * 28).to(device)
        Y = Y.to(device)

        optimizer.zero_grad()
        hypothesis = linear(X)
        cost = criterion(hypothesis, Y)
        cost.backward()
        optimizer.step()

        avg_cost += cost / total_batch

    print('Epoch:', '%04d' % (epoch + 1), 'cost =', '{:.9f}'.format(avg_cost))

print('Learning finished')

'''output
Epoch: 0001 cost = 4.848181248
Epoch: 0002 cost = 1.464641452
Epoch: 0003 cost = 0.977406502
Epoch: 0004 cost = 0.790303528
Epoch: 0005 cost = 0.686833322
Epoch: 0006 cost = 0.618483305
Epoch: 0007 cost = 0.568978667
Epoch: 0008 cost = 0.531290889
Epoch: 0009 cost = 0.501056492
Epoch: 0010 cost = 0.476258427
Epoch: 0011 cost = 0.455025405
Epoch: 0012 cost = 0.437031567
Epoch: 0013 cost = 0.421489984
Epoch: 0014 cost = 0.408599794
Epoch: 0015 cost = 0.396514893
Learning finished
'''