Wejdz na stronę codebook. Zarejestruj się i zaloguj.
IQC: Introduction to Quantum Computing
Wykonaj pierwsze trzy ćwiczenia z sekcji wprowadzenie od obliczeń kwantowych.
\[
\newcommand{\bra}[1]{\left \langle #1 \right \rvert}
\newcommand{\ket}[1]{\left \rvert #1 \right \rangle}
\newcommand{\braket}[2]{\left \langle #1 \middle \rvert #2 \right \rangle}
\]
PyTorch Regresja liniowa
import torch
import numpy as np
import matplotlib.pyplot as plt
class LinearRegression(torch.nn.Module):
def __init__(self, inputSize, outputSize):
super(LinearRegression, self).__init__()
self.linear = torch.nn.Linear(inputSize, outputSize)
def forward(self, x):
out = self.linear(x)
return out
inputDim = 1
outputDim = 1
learningRate = 0.01
epochs = 100
model = LinearRegression(inputDim, outputDim)
x = np.arange(11)
print(f"dane: {x}")
y = [2*el - 1 for el in x]
print(f"target: {y}")
# dostosowanie do pytorch
y = np.array(y, dtype=np.float32)
x_train = np.array(x, dtype=np.float32)
x_train, y_train = x_train.reshape(-1, 1), y.reshape(-1, 1)
inputs = torch.from_numpy(x_train)
outputs = torch.from_numpy(y_train)
criterion = torch.nn.MSELoss()
optimizer = torch.optim.SGD(model.parameters(), lr=learningRate)
# petla uczaca
for epoch in range(epochs):
# forward pass and loss
y_predicted = model(inputs)
loss = criterion(y_predicted, outputs)
# backward pass
loss.backward()
# updates
optimizer.step()
# zero gradients
optimizer.zero_grad()
if (epoch+1) % 10 == 0:
print(f'epoch: {epoch+1}, loss = {loss.item():.4f}')
with torch.no_grad():
predicted = model(inputs)
plt.clf()
plt.plot(x_train, y_train, 'go', label='True data', alpha=0.5)
plt.plot(x_train, predicted, '--', label='Predictions', alpha=0.5)
plt.legend(loc='best')
plt.show()
dane: [ 0 1 2 3 4 5 6 7 8 9 10]
target: [-1, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19]
epoch: 10, loss = 0.2569
epoch: 20, loss = 0.2296
epoch: 30, loss = 0.2052
epoch: 40, loss = 0.1834
epoch: 50, loss = 0.1640
epoch: 60, loss = 0.1465
epoch: 70, loss = 0.1310
epoch: 80, loss = 0.1171
epoch: 90, loss = 0.1046
epoch: 100, loss = 0.0935
PyTorch Regresja logistyczna
import torch
import numpy as np
import matplotlib.pyplot as plt
class LogisticRegression(torch.nn.Module):
def __init__(self, inputSize, outputSize):
super(LogisticRegression, self).__init__()
self.linear = torch.nn.Linear(inputSize, outputSize)
def forward(self, x):
x = self.linear(x)
y_pred = torch.sigmoid(x)
return y_pred
inputDim = 1
outputDim = 1
learningRate = 0.001
epochs = 100
model = LogisticRegression(inputDim, outputDim)
x = np.arange(11)
print(f"dane: {x}")
import random
y = [random.choice([0,1]) for _ in range(11)]
print(f"target: {y}")
# dostosowanie do pytorch
y = np.array(y, dtype=np.float32)
x_train = np.array(x, dtype=np.float32)
x_train, y_train = x_train.reshape(-1, 1), y.reshape(-1, 1)
inputs = torch.from_numpy(x_train)
outputs = torch.from_numpy(y_train)
criterion = torch.nn.CrossEntropyLoss()
optimizer = torch.optim.SGD(model.parameters(), lr=learningRate)
# petla uczaca
for epoch in range(epochs):
# forward pass and loss
y_predicted = model(inputs)
loss = criterion(y_predicted, outputs)
# backward pass
loss.backward()
# updates
optimizer.step()
# zero gradients
optimizer.zero_grad()
if (epoch+1) % 10 == 0:
print(f'epoch: {epoch+1}, loss = {loss.item():.4f}')
with torch.no_grad():
y_predicted = model(inputs)
acc = y_predicted.eq(outputs).sum() / float(y_train.shape[0]) # accuracy
print(f'accuracy = {acc:.4f}')
dane: [ 0 1 2 3 4 5 6 7 8 9 10]
target: [0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0]
epoch: 10, loss = -0.0000
epoch: 20, loss = -0.0000
epoch: 30, loss = -0.0000
epoch: 40, loss = -0.0000
epoch: 50, loss = -0.0000
epoch: 60, loss = -0.0000
epoch: 70, loss = -0.0000
epoch: 80, loss = -0.0000
epoch: 90, loss = -0.0000
epoch: 100, loss = -0.0000
accuracy = 0.0000
