Tag: overfitting

model.train()    # set the model to train mode (dropout=True)
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 = model(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 = 0.308392197
Epoch: 0002 cost = 0.142623395
Epoch: 0003 cost = 0.113427199
Epoch: 0004 cost = 0.093490042
Epoch: 0005 cost = 0.083772294
Epoch: 0006 cost = 0.077040948
Epoch: 0007 cost = 0.067025252
Epoch: 0008 cost = 0.063156039
Epoch: 0009 cost = 0.058766391
Epoch: 0010 cost = 0.055902217
Epoch: 0011 cost = 0.052059878
Epoch: 0012 cost = 0.048243146
Epoch: 0013 cost = 0.047231019
Epoch: 0014 cost = 0.045120358
Epoch: 0015 cost = 0.040942233
Learning finished
'''

model = SoftmaxClassifierModel()
optimizer = optim.SGD(model.parameters(), lr=1e5)
train(model, optimizer, x_train, y_train)

'''output
Epoch    0/20 Cost: 1.280268
Epoch    1/20 Cost: 976950.812500
Epoch    2/20 Cost: 1279135.125000
Epoch    3/20 Cost: 1198379.000000
Epoch    4/20 Cost: 1098825.875000
Epoch    5/20 Cost: 1968197.625000
Epoch    6/20 Cost: 284763.250000
Epoch    7/20 Cost: 1532260.125000
Epoch    8/20 Cost: 1651504.000000
Epoch    9/20 Cost: 521878.500000
Epoch   10/20 Cost: 1397263.250000
Epoch   11/20 Cost: 750986.250000
Epoch   12/20 Cost: 918691.500000
Epoch   13/20 Cost: 1487888.250000
Epoch   14/20 Cost: 1582260.125000
Epoch   15/20 Cost: 685818.062500
Epoch   16/20 Cost: 1140048.750000
Epoch   17/20 Cost: 940566.500000
Epoch   18/20 Cost: 931638.250000
Epoch   19/20 Cost: 1971322.625000
'''