from libdnn import Classifier
# multi layer perceptron
model = chainer.FunctionSet(l1=F.Linear(2, 100), l2=F.Linear(100, 100), l3=F.Linear(100, 2))
# define forwarding method
def forward(self, x, train):
h = F.relu(self.model.l1(x))
h = F.relu(self.model.l2(h))
y = self.model.l3(h)
return y
mlp = Classifier(model, gpu=-1)
mlp.set_forward(forward)
mlp.set_optimizer(Opt.AdaDelta, {'rho': 0.9})
arr = []
t = []
for i in range(10000):
x, y = (numpy.random.rand() - 0.5), (numpy.random.rand() - 0.5)
arr.append(numpy.array([x, y]))
if (x < 0. and y < 0.) or (x > 0. and y > 0.):
t.append(0)
else:
t.append(1)
print(mlp.train(numpy.array(arr).astype(numpy.float32), numpy.array(t).astype(numpy.int32)))
print(mlp.test(numpy.array(arr).astype(numpy.float32), numpy.array(t).astype(numpy.int32)))
# multi layer perceptron
model = chainer.FunctionSet(l1=F.Linear(2, 100),
l2=F.Linear(100, 100),
l3=F.Linear(100, 2))
# define forwarding method
def forward(self, x, train):
h = F.relu(self.model.l1(x))
h = F.relu(self.model.l2(h))
y = self.model.l3(h)
return y
mlp = Classifier(model, gpu=-1)
mlp.set_forward(forward)
mlp.set_optimizer(Opt.AdaDelta, {'rho': 0.9})
arr = []
t = []
for i in range(10000):
x, y = (numpy.random.rand() - 0.5), (numpy.random.rand() - 0.5)
arr.append(numpy.array([x, y]))
if (x < 0. and y < 0.) or (x > 0. and y > 0.):
t.append(0)
else:
t.append(1)
print(
mlp.train(
fl4=F.Linear(2304, 576),
fl5=F.Linear(576, 10)
)
def forward(self, x, train):
h = F.max_pooling_2d(F.relu(model.bn1(model.conv1(x))), 2)
h = F.relu(model.bn2(model.conv2(h)))
h = F.max_pooling_2d(F.relu(model.conv3(h)), 2)
h = F.dropout(F.relu(model.fl4(h)), train=False)
y = model.fl5(h)
return y
cnn = Classifier(model, gpu=-1)
cnn.set_forward(forward)
mnist = fetch_mldata('MNIST original', data_home='.')
perm = numpy.random.permutation(len(mnist.data))
mnist.data = mnist.data.astype(numpy.float32).reshape(70000, 1, 28, 28) / 255
mnist.target = mnist.target.astype(numpy.int32)
train_data = mnist.data[perm][:60000]
train_label = mnist.target[perm][:60000]
test_data = mnist.data[perm][60000:]
test_label = mnist.target[perm][60000:]
for epoch in range(15):
print('epoch : %d' % (epoch + 1))
err, acc = cnn.train(train_data, train_label, batchsize=200)
print(acc, err)
conv3=F.Convolution2D(30, 64, 3, pad=1),
fl4=F.Linear(2304, 576),
fl5=F.Linear(576, 10))
def forward(self, x, train):
h = F.max_pooling_2d(F.relu(model.bn1(model.conv1(x))), 2)
h = F.relu(model.bn2(model.conv2(h)))
h = F.max_pooling_2d(F.relu(model.conv3(h)), 2)
h = F.dropout(F.relu(model.fl4(h)), train=False)
y = model.fl5(h)
return y
cnn = Classifier(model, gpu=-1)
cnn.set_forward(forward)
mnist = fetch_mldata('MNIST original', data_home='.')
perm = numpy.random.permutation(len(mnist.data))
mnist.data = mnist.data.astype(numpy.float32).reshape(70000, 1, 28, 28) / 255
mnist.target = mnist.target.astype(numpy.int32)
train_data = mnist.data[perm][:60000]
train_label = mnist.target[perm][:60000]
test_data = mnist.data[perm][60000:]
test_label = mnist.target[perm][60000:]
for epoch in range(15):
print('epoch : %d' % (epoch + 1))
err, acc = cnn.train(train_data, train_label, batchsize=200)
print(acc, err)
fl4=F.Linear(576, 576),
fl5=F.Linear(576, 10))
def forward(self, x, train):
h = F.max_pooling_2d(F.relu(model.bn1(model.conv1(x))), 2)
h = F.max_pooling_2d(h, 2)
h = F.max_pooling_2d(F.relu(model.bn2(model.conv2(h))), 2)
h = F.max_pooling_2d(F.relu(model.conv3(h)), 2)
h = F.dropout(F.relu(model.fl4(h)), train=True)
y = model.fl5(h)
return y
cnn = Classifier(model, gpu=-1)
cnn.set_forward(forward)
arr = []
t = []
for i in range(100):
x = numpy.random.rand(50, 50)
x = numpy.array([x])
arr.append(x)
a = numpy.random.randint(0, 9)
t.append(a)
print(
cnn.train(
numpy.array(arr).astype(numpy.float32),
numpy.array(t).astype(numpy.int32)))
fl4=F.Linear(576, 576),
fl5=F.Linear(576, 10)
)
def forward(self, x, train):
h = F.max_pooling_2d(F.relu(model.bn1(model.conv1(x))), 2)
h = F.max_pooling_2d(h, 2)
h = F.max_pooling_2d(F.relu(model.bn2(model.conv2(h))), 2)
h = F.max_pooling_2d(F.relu(model.conv3(h)), 2)
h = F.dropout(F.relu(model.fl4(h)), train=True)
y = model.fl5(h)
return y
cnn = Classifier(model, gpu=-1)
cnn.set_forward(forward)
arr = []
t = []
for i in range(100):
x = numpy.random.rand(50, 50)
x = numpy.array([x])
arr.append(x)
a = numpy.random.randint(0, 9)
t.append(a)
print(cnn.train(numpy.array(arr).astype(numpy.float32), numpy.array(t).astype(numpy.int32)))
print(cnn.test(numpy.array(arr).astype(numpy.float32), numpy.array(t).astype(numpy.int32)))
model = chainer.FunctionSet(
conv1=F.Convolution2D(1, 15, 5),
bn1=F.BatchNormalization(15),
conv2=F.Convolution2D(15, 30, 3, pad=1),
bn2=F.BatchNormalization(30),
conv3=F.Convolution2D(30, 64, 3, pad=1),
fl4=F.Linear(2304, 576),
fl5=F.Linear(576, 10)
)
def forward(self, x, train):
h = F.max_pooling_2d(F.relu(model.bn1(model.conv1(x))), 2)
h = F.relu(model.bn2(model.conv2(h)))
h = F.max_pooling_2d(F.relu(model.conv3(h)), 2)
h = F.dropout(F.relu(model.fl4(h)), train=True)
y = model.fl5(h)
return y
cnn = Classifier(model, gpu=-1)
cnn.set_forward(forward)
cnn.load_param('./cnn.param.npy')
imager = V.Visualizer(cnn)
imager.plot_filters('conv1', interpolation=True)
imager.plot_filters('conv2', title=False, interpolation=True)
imager.plot_filters('conv3', title=False, interpolation=True)