08-1 perceptron

목차

• erceptron
• ND, OR
• OR
• Code: XOR

# Lab 9 XOR
import torch

device = 'cuda' if torch.cuda.is_available() else 'cpu'

# for reproducibility
torch.manual_seed(777)
if device == 'cuda':
    torch.cuda.manual_seed_all(777)

X = torch.FloatTensor([[0, 0], [0, 1], [1, 0], [1, 1]]).to(device) # XOR data
Y = torch.FloatTensor([[0], [1], [1], [0]]).to(device)

# nn layers
linear = torch.nn.Linear(2, 1, bias=True)
sigmoid = torch.nn.Sigmoid()

# model
model = torch.nn.Sequential(linear, sigmoid).to(device)

# define cost/loss & optimizer
criterion = torch.nn.BCELoss().to(device)
optimizer = torch.optim.SGD(model.parameters(), lr=1)

for step in range(10001):
    optimizer.zero_grad()
    hypothesis = model(X)

    # cost/loss function
    cost = criterion(hypothesis, Y)
    cost.backward()
    optimizer.step()

    if step % 100 == 0:
        print(step, cost.item())

0 0.7273974418640137
100 0.6931475400924683
200 0.6931471824645996
300 0.6931471824645996
400 0.6931471824645996
500 0.6931471824645996
600 0.6931471824645996
700 0.6931471824645996
800 0.6931471824645996
900 0.6931471824645996
1000 0.6931471824645996
1100 0.6931471824645996
1200 0.6931471824645996
1300 0.6931471824645996
1400 0.6931471824645996
1500 0.6931471824645996
1600 0.6931471824645996
1700 0.6931471824645996
1800 0.6931471824645996
1900 0.6931471824645996
2000 0.6931471824645996
2100 0.6931471824645996
2200 0.6931471824645996
2300 0.6931471824645996
2400 0.6931471824645996
2500 0.6931471824645996
2600 0.6931471824645996
2700 0.6931471824645996
2800 0.6931471824645996
2900 0.6931471824645996
3000 0.6931471824645996
3100 0.6931471824645996
3200 0.6931471824645996
3300 0.6931471824645996
3400 0.6931471824645996
3500 0.6931471824645996
3600 0.6931471824645996
3700 0.6931471824645996
3800 0.6931471824645996
3900 0.6931471824645996
4000 0.6931471824645996
4100 0.6931471824645996
4200 0.6931471824645996
4300 0.6931471824645996
4400 0.6931471824645996
4500 0.6931471824645996
4600 0.6931471824645996
4700 0.6931471824645996
4800 0.6931471824645996
4900 0.6931471824645996
5000 0.6931471824645996
5100 0.6931471824645996
5200 0.6931471824645996
5300 0.6931471824645996
5400 0.6931471824645996
5500 0.6931471824645996
5600 0.6931471824645996
5700 0.6931471824645996
5800 0.6931471824645996
5900 0.6931471824645996
6000 0.6931471824645996
6100 0.6931471824645996
6200 0.6931471824645996
6300 0.6931471824645996
6400 0.6931471824645996
6500 0.6931471824645996
6600 0.6931471824645996
6700 0.6931471824645996
6800 0.6931471824645996
6900 0.6931471824645996
7000 0.6931471824645996
7100 0.6931471824645996
7200 0.6931471824645996
7300 0.6931471824645996
7400 0.6931471824645996
7500 0.6931471824645996
7600 0.6931471824645996
7700 0.6931471824645996
7800 0.6931471824645996
7900 0.6931471824645996
8000 0.6931471824645996
8100 0.6931471824645996
8200 0.6931471824645996
8300 0.6931471824645996
8400 0.6931471824645996
8500 0.6931471824645996
8600 0.6931471824645996
8700 0.6931471824645996
8800 0.6931471824645996
8900 0.6931471824645996
9000 0.6931471824645996
9100 0.6931471824645996
9200 0.6931471824645996
9300 0.6931471824645996
9400 0.6931471824645996
9500 0.6931471824645996
9600 0.6931471824645996
9700 0.6931471824645996
9800 0.6931471824645996
9900 0.6931471824645996
10000 0.6931471824645996

# Accuracy computation
# True if hypothesis>0.5 else False
with torch.no_grad():
    hypothesis = model(X)
    predicted = (hypothesis > 0.5).float()
    accuracy = (predicted == Y).float().mean()
    print('\nHypothesis: ', hypothesis.detach().cpu().numpy(), '\nCorrect: ', predicted.detach().cpu().numpy(), '\nAccuracy: ', accuracy.item())

Hypothesis:  [[0.5]
 [0.5]
 [0.5]
 [0.5]] 
Correct:  [[0.]
 [0.]
 [0.]
 [0.]] 
Accuracy:  0.5