# layer_0
model_0 = ConvAutoEncoder(
layers=[layer_0_en,
MaxPoolingSameSize(pool_size=(4, 4)), layer_0_de])
out_0 = model_0.fprop(images, corruption_level=corruption_level)
cost_0 = mean_square_cost(out_0[-1], images) + L2_regularization(
model_0.params, 0.005)
updates_0 = gd_updates(cost=cost_0,
params=model_0.params,
method="sgd",
learning_rate=0.1)
# layer_0 --> layer_1
model_0_to_1 = FeedForward(
layers=[layer_0_en, MaxPooling(pool_size=(4, 4))])
out_0_to_1 = model_0_to_1.fprop(images)
# layer_1
model_1 = ConvAutoEncoder(
layers=[layer_1_en,
MaxPoolingSameSize(pool_size=(2, 2)), layer_1_de])
out_1 = model_1.fprop(out_0_to_1[-1], corruption_level=corruption_level)
cost_1 = mean_square_cost(out_1[-1], out_0_to_1[-1]) + L2_regularization(
model_1.params, 0.005)
updates_1 = gd_updates(cost=cost_1,
params=model_1.params,
method="sgd",
learning_rate=0.1)
# layer_1 --> layer_2
model_1_to_2 = FeedForward(
print ' ' , np.mean(c_2), str(corr_best[2][0]), min_cost[2], max_iter[2]
print ' ' , np.mean(c_3), str(corr_best[3][0]), min_cost[3], max_iter[3]
print "[MESSAGE] The model is trained"
################################## BUILD SUPERVISED MODEL #######################################
flattener=Flattener()
layer_5=ReLULayer(in_dim=50*16*16,
out_dim=1000)
layer_6=SoftmaxLayer(in_dim=1000,
out_dim=10)
model_sup=FeedForward(layers=[layer_0_en, layer_1_en, layer_2_en, layer_3_en, flattener, layer_5, layer_6])
out_sup=model_sup.fprop(images)
cost_sup=categorical_cross_entropy_cost(out_sup[-1], y)
updates=gd_updates(cost=cost_sup, params=model_sup.params, method="sgd", learning_rate=0.1)
train_sup=theano.function(inputs=[idx],
outputs=cost_sup,
updates=updates,
givens={X: train_set_x[idx * batch_size: (idx + 1) * batch_size],
y: train_set_y[idx * batch_size: (idx + 1) * batch_size]})
test_sup=theano.function(inputs=[idx],
outputs=model_sup.layers[-1].error(out_sup[-1], y),
givens={X: test_set_x[idx * batch_size: (idx + 1) * batch_size],
y: test_set_y[idx * batch_size: (idx + 1) * batch_size]})
print "[MESSAGE] The supervised model is built"
border_mode="same") # learning rate formula: # r = 1 - 0.5*(ni/ntot)*(ni/ntot) # ni = ith layer; ntot = number of layers # layer_0 model_0=ConvAutoEncoder(layers=[layer_0_en, MaxPoolingSameSize(pool_size=(4,4)), layer_0_de]) out_0=model_0.fprop(images, corruption_level=corruption_level) cost_0=mean_square_cost(out_0[-1], images)+L2_regularization(model_0.params, 0.005) updates_0=gd_updates(cost=cost_0, params=model_0.params, method="sgd", learning_rate=0.1) # layer_0 --> layer_1 model_0_to_1=FeedForward(layers=[layer_0_en, MaxPooling(pool_size=(2,2))]); out_0_to_1=model_0_to_1.fprop(images); # layer_1 model_1=ConvAutoEncoder(layers=[layer_1_en, MaxPoolingSameSize(pool_size=(2,2)), layer_1_de]) out_1=model_1.fprop(out_0_to_1[-1], corruption_level=corruption_level) cost_1=mean_square_cost(out_1[-1], out_0_to_1[-1])+L2_regularization(model_1.params, 0.005) updates_1=gd_updates(cost=cost_1, params=model_1.params, method="sgd", learning_rate=0.1) # layer_1 --> layer_2 model_1_to_2=FeedForward(layers=[layer_1_en, MaxPooling(pool_size=(4,4))]); out_1_to_2=model_1_to_2.fprop(images); # layer_2 model_2=ConvAutoEncoder(layers=[layer_2_en, MaxPoolingSameSize(pool_size=(2,2)), layer_2_de]) out_2=model_2.fprop(out_1_to_2[-1], corruption_level=corruption_level) cost_2=mean_square_cost(out_2[-1], out_1_to_2[-1])+L2_regularization(model_2.params, 0.005)
# layer_3_de=SigmoidConvLayer(filter_size=(3,3), # num_filters=50, # num_channels=50, # fm_size=(16,16), # batch_size=batch_size, # border_mode="full") model_0=ConvAutoEncoder(layers=[layer_0_en, layer_0_de]) out_0=model_0.fprop(images, corruption_level=corruption_level) cost_0=mean_square_cost(out_0[-1], images)+L2_regularization(model_0.params, 0.005) updates_0=gd_updates(cost=cost_0, params=model_0.params, method="sgd", learning_rate=0.1) ## append a max-pooling layer model_trans=FeedForward(layers=[layer_0_en, MaxPooling(pool_size=(2,2))]); out_trans=model_trans.fprop(images); model_1=ConvAutoEncoder(layers=[layer_1_en, layer_1_de]) out_1=model_1.fprop(out_trans[-1], corruption_level=corruption_level) cost_1=mean_square_cost(out_1[-1], out_trans[-1])+L2_regularization(model_1.params, 0.005) updates_1=gd_updates(cost=cost_1, params=model_1.params, method="sgd", learning_rate=0.1) # model_2=ConvAutoEncoder(layers=[layer_2_en, layer_2_de]) # out_2=model_2.fprop(out_1[0], corruption_level=corruption_level) # cost_2=mean_square_cost(out_2[-1], out_1[0])+L2_regularization(model_2.params, 0.005) # updates_2=gd_updates(cost=cost_2, params=model_2.params, method="sgd", learning_rate=0.1) # model_3=ConvAutoEncoder(layers=[layer_3_en, layer_3_de]) # out_3=model_3.fprop(out_2[0], corruption_level=corruption_level) # cost_3=mean_square_cost(out_3[-1], out_2[0])+L2_regularization(model_3.params, 0.005)
# border_mode="full")
model_0 = ConvAutoEncoder(layers=[layer_0_en, layer_0_de])
out_0 = model_0.fprop(images, corruption_level=corruption_level)
cost_0 = mean_square_cost(out_0[-1], images) + L2_regularization(
model_0.params, 0.005)
updates_0 = gd_updates(cost=cost_0,
params=model_0.params,
method="sgd",
learning_rate=0.1)
## append a max-pooling layer
model_trans = FeedForward(
layers=[layer_0_en, MaxPooling(pool_size=(2, 2))])
out_trans = model_trans.fprop(images)
model_1 = ConvAutoEncoder(layers=[layer_1_en, layer_1_de])
out_1 = model_1.fprop(out_trans[-1], corruption_level=corruption_level)
cost_1 = mean_square_cost(out_1[-1], out_trans[-1]) + L2_regularization(
model_1.params, 0.005)
updates_1 = gd_updates(cost=cost_1,
params=model_1.params,
method="sgd",
learning_rate=0.1)
# model_2=ConvAutoEncoder(layers=[layer_2_en, layer_2_de])
# out_2=model_2.fprop(out_1[0], corruption_level=corruption_level)
# cost_2=mean_square_cost(out_2[-1], out_1[0])+L2_regularization(model_2.params, 0.005)
# updates_2=gd_updates(cost=cost_2, params=model_2.params, method="sgd", learning_rate=0.1)
fm_size=(16, 16),
batch_size=batch_size,
border_mode="full")
pool_1 = MaxPooling(pool_size=(2, 2))
flattener = Flattener()
layer_2 = ReLULayer(in_dim=32 * 64, out_dim=800)
layer_3 = SoftmaxLayer(in_dim=800, out_dim=10)
model = FeedForward(
layers=[layer_0, pool_0, layer_1, pool_1, flattener, layer_2, layer_3])
out = model.fprop(images)
cost = categorical_cross_entropy_cost(out[-1], y)
updates = gd_updates(cost=cost,
params=model.params,
method="sgd",
learning_rate=0.01,
momentum=0.9)
extract = theano.function(
inputs=[idx],
outputs=layer_0.apply(images),
givens={X: train_set_x[idx * batch_size:(idx + 1) * batch_size]})
print extract(1).shape
train = theano.function(
inputs=[idx],
batch_size=batch_size,
border_mode="full");
pool_1=MaxPooling(pool_size=(2,2));
flattener=Flattener();
layer_2=ReLULayer(in_dim=32*64,
out_dim=800);
layer_3=SoftmaxLayer(in_dim=800,
out_dim=10);
model=FeedForward(layers=[layer_0, pool_0, layer_1, pool_1, flattener, layer_2, layer_3]);
out=model.fprop(images);
cost=categorical_cross_entropy_cost(out[-1], y);
updates=gd_updates(cost=cost, params=model.params, method="sgd", learning_rate=0.01, momentum=0.9);
extract=theano.function(inputs=[idx],
outputs=layer_0.apply(images),
givens={X: train_set_x[idx * batch_size: (idx + 1) * batch_size]});
print extract(1).shape
train=theano.function(inputs=[idx],
outputs=cost,
updates=updates,
givens={X: train_set_x[idx * batch_size: (idx + 1) * batch_size],
y: train_set_y[idx * batch_size: (idx + 1) * batch_size]});
print ' ', np.mean(c_3), str(
corr_best[3][0]), min_cost[3], max_iter[3]
print "[MESSAGE] The model is trained"
################################## BUILD SUPERVISED MODEL #######################################
flattener = Flattener()
layer_5 = ReLULayer(in_dim=50 * 16 * 16, out_dim=1000)
layer_6 = SoftmaxLayer(in_dim=1000, out_dim=10)
model_sup = FeedForward(layers=[
layer_0_en, layer_1_en, layer_2_en, layer_3_en, flattener, layer_5, layer_6
])
out_sup = model_sup.fprop(images)
cost_sup = categorical_cross_entropy_cost(out_sup[-1], y)
updates = gd_updates(cost=cost_sup,
params=model_sup.params,
method="sgd",
learning_rate=0.1)
train_sup = theano.function(
inputs=[idx],
outputs=cost_sup,
updates=updates,
givens={
X: train_set_x[idx * batch_size:(idx + 1) * batch_size],
y: train_set_y[idx * batch_size:(idx + 1) * batch_size]
})