class TwoLayerNet(Model):
def __init__(self, in_size, hidden_size, out_size, activation=F.sigmoid):
super().__init__()
self.f = activation
self.l1 = L.Linear(in_size, hidden_size)
self.l2 = L.Linear(hidden_size, out_size)
def __call__(self, x):
y = self.f(self.l1(x))
y = self.l2(y)
return y
model = TwoLayerNet(784, hidden_size, 10)
optimizer = optimizers.SGD().setup(model)
for epoch in range(max_epoch):
sum_loss, sum_acc = 0, 0
for x, t in train_loader:
y = model(x)
loss = F.softmax_cross_entropy(y, t)
acc = F.accuracy(y, t)
model.cleargrads()
loss.backward()
optimizer.update()
sum_loss += float(loss.data) * len(t)
sum_acc += float(acc.data) * len(t)
import numpy as np
import matplotlib.pyplot as plt
import dezero
from dezero import optimizers
import dezero.functions as F
from dezero.datasets import get_spiral
from dezero.models import TwoLayerNet
# ハイパーパラメータの設定
max_epoch = 300
batch_size = 30
hidden_size = 10
lr = 1.0
x, t = get_spiral()
model = TwoLayerNet(2, hidden_size, 3)
optimizer = optimizers.SGD(lr).setup(model)
data_size = len(x)
max_iter = math.ceil(data_size / batch_size) # 小数点の切り上げ
loss_list = []
for epoch in range(max_epoch):
# データのシャッフル
idx = np.random.permutation(data_size)
x = x[idx]
t = t[idx]
sum_loss = 0
for i in range(max_iter):
import dezero
import dezero.functions as F
from dezero import optimizers
from dezero.dataset import DatasetLoader
from dezero.models import TwoLayerNet
max_epoch = 3
batch_size = 100
train_set, test_set = dezero.datasets.get_mnist()
train_loader = DatasetLoader(train_set, batch_size)
model = TwoLayerNet(784, 1000, 10)
optimizer = optimizers.SGD().setup(model)
# パラーメタの読み込み
model.load_weights('two_layer_net.npz')
for epoch in range(max_epoch):
sum_loss = 0
for x, t in train_loader:
y = model(x)
loss = F.softmax_cross_entropy(y, t)
model.cleargrads()
loss.backward()
optimizer.update()
sum_loss += float(loss.data) * len(t)
print('epoch: {}, loss: {:.4f}'.format(epoch + 1,
sum_loss / len(train_set)))
from dezero import optimizers
import dezero.functions as F
from dezero.models import TwoLayerNet
np.random.seed(0)
x = np.random.rand(100, 1)
y = np.sin(2 * np.pi * x) + np.random.rand(100, 1)
x, y = Variable(x), Variable(y)
lr = 0.2
max_iter = 10000
hidden_size = 10
N, I = x.shape
N, O = y.shape
model = TwoLayerNet(I, hidden_size, O)
optimizer = optimizers.SGD(lr).setup(model)
for i in range(max_iter):
y_pred = model(x)
loss = F.mean_squared_error(y, y_pred)
model.cleargrads()
loss.backward()
optimizer.update()
print(loss)
# グラフの描画
import matplotlib.pyplot as plt
from dezero import Variable
import dezero.functions as F
from dezero.models import TwoLayerNet
# データセットの生成
np.random.seed(0)
x = np.random.rand(100, 1)
y = np.sin(2 * np.pi * x) + np.random.rand(100, 1)
x, y = Variable(x), Variable(y)
# ハイパーパラメータの設定
lr = 0.2
max_iter = 10000
hidden_size = 10
# モデルの生成
N, I = x.shape
N, O = y.shape
model = TwoLayerNet(I, hidden_size, O)
# 学習の開始
for i in range(max_iter):
y_pred = model(x)
loss = F.mean_squared_error(y, y_pred)
model.cleargrads()
loss.backward()
for p in model.params():
p.data -= lr * p.grad.data
print(loss)
import time
import dezero
import dezero.functions as F
from dezero import optimizers
from dezero.dataset import DatasetLoader
from dezero.models import TwoLayerNet
max_epoch = 3
batch_size = 100
train_set, test_set = dezero.datasets.get_mnist()
train_loader = DatasetLoader(train_set, batch_size)
model = TwoLayerNet(784, 1000, 10)
optimizer = optimizers.SGD().setup(model)
# GPU mode
train_loader.to_gpu()
model.to_gpu()
for epoch in range(max_epoch):
start = time.time()
sum_loss = 0
for x, t in train_loader:
y = model(x)
loss = F.softmax_cross_entropy(y, t)
model.cleargrads()
loss.backward()
optimizer.update()
sum_loss += float(loss.data) * len(t)
import numpy as np
np.random.seed(0) # シードを固定
from dezero import Variable, Model
import dezero.functions as F
from dezero.models import TwoLayerNet
def softmax1d(x):
y = F.exp(x)
sum_y = F.sum(y)
return y / sum_y
model = TwoLayerNet(in_size=2, hidden_size=10, out_size=3)
x = Variable(np.array([0.2, -0.4]))
y = model(x)
p = softmax1d(y)
print(y) # variable([-0.6150578 -0.42790162 0.31733288])
print(p) # variable([0.21068638 0.25404893 0.53526469])
x = np.array([[0.2, -0.4], [0.3, 0.5], [1.3, -3.2], [2.1, 0.3]])
y = model(x)
p = F.softmax(y)
print(y)
# variable([[-0.6150578 -0.42790162 0.31733288]
# [-0.76395315 -0.24976449 0.18591381]
# [-0.52006394 -0.96254613 0.57818937]
# [-0.94252168 -0.50307479 0.17576322]])