def all_classifier(feature1, feature2, keep_prob):
with tf.name_scope('all_classifier'):
W1 = utils.get_weights_variable(4096, 2048, name='W1')
b1 = utils.get_bias_variable(2048, name='b1')
matmul1 = tf.nn.bias_add(tf.matmul(feature1, W1), b1)
matmul1 = tf.nn.relu(matmul1)
W2 = utils.get_weights_variable(4096, 2048, name='W2')
b2 = utils.get_bias_variable(2048, name='b2')
matmul2 = tf.nn.bias_add(tf.matmul(feature2, W2), b2)
matmul2 = tf.nn.relu(matmul2)
feature = tf.concat([matmul1, matmul2], axis=1)
feature = tf.nn.dropout(feature, keep_prob)
W3 = utils.get_weights_variable(4096, 4096, name='W3')
b3 = utils.get_bias_variable(4096, name='b3')
matmul3 = tf.nn.bias_add(tf.matmul(feature, W3), b3)
matmul3 = tf.nn.relu(matmul3)
W5 = utils.get_weights_variable(4096, CLASS_COUNTS, name='W5')
b5 = utils.get_bias_variable(CLASS_COUNTS, name='b5')
matmul5 = tf.nn.bias_add(tf.matmul(matmul3, W5), b5)
return matmul5
def all_classifier(feature1, feature2, keep_prob):
with tf.name_scope('all_classifier'):
W1 = utils.get_weights_variable(4096, 3072, name='W1')
b1 = utils.get_bias_variable(3072, name='b1')
matmul1 = tf.nn.bias_add(tf.matmul(feature1, W1), b1)
matmul1 = tf.nn.relu(matmul1)
W2 = utils.get_weights_variable(4096, 1024, name='W2')
b2 = utils.get_bias_variable(1024, name='b2')
matmul2 = tf.nn.bias_add(tf.matmul(feature2, W2), b2)
matmul2 = tf.nn.relu(matmul2)
feature = tf.concat([matmul1, matmul2], axis=1)
feature = tf.nn.dropout(feature, keep_prob)
W3 = utils.get_weights_variable(4096, 3072, name='W3')
b3 = utils.get_bias_variable(3072, name='b3')
matmul3 = tf.nn.bias_add(tf.matmul(feature, W3), b3)
matmul3 = tf.nn.relu(matmul3)
matmul3 = tf.concat([matmul3, matmul2], axis=1)
matmul3 = tf.nn.dropout(matmul3, keep_prob)
W4 = utils.get_weights_variable(4096, CLASS_COUNTS, name='W4')
b4 = utils.get_bias_variable(CLASS_COUNTS, name='b4')
matmul4 = tf.nn.bias_add(tf.matmul(matmul3, W4), b4)
return matmul4
def subnet1(image, keep_prob, reuse=False):
with tf.variable_scope("subnet1", reuse=reuse):
def _vgg_net(weights, image):
print('setting up vgg model initialized params --> extractor1')
layers = (
'conv1_1', 'relu1_1', 'conv1_2', 'relu1_2', 'pool1',
'conv2_1', 'relu2_1', 'conv2_2', 'relu2_2', 'pool2',
'conv3_1', 'relu3_1', 'conv3_2', 'relu3_2', 'conv3_3',
'relu3_3', 'conv3_4', 'relu3_4', 'pool3',
'conv4_1', 'relu4_1', 'conv4_2', 'relu4_2', 'conv4_3',
'relu4_3', 'conv4_4', 'relu4_4', 'pool4',
'conv5_1', 'relu5_1', 'conv5_2', 'relu5_2', 'conv5_3',
'relu5_3', 'conv5_4', 'relu5_4'
)
net = {}
current = image
for i, name in enumerate(layers):
kind = name[:4]
if kind == 'conv':
kernels, bias = weights[i][0][0][0][0]
# kernels are [width, height, in_channles, out_channles]
# tensorflow are [height, width, in channles, out_channles]
kernels = utils.get_variable(
np.transpose(kernels, (1, 0, 2, 3)), name=name + '_w')
bias = utils.get_variable(bias.reshape(-1), name=name + '_b')
current = utils.conv2d_basic(current, kernels, bias)
elif kind == 'relu':
current = tf.nn.relu(current, name=name)
elif kind == 'pool':
current = utils.avg_pool_2x2(current)
net[name] = current
return net
model_data = utils.get_model_data("data", MODEL_URL)
weights = np.squeeze(model_data['layers'])
with tf.variable_scope('inference'):
image_net = _vgg_net(weights, image)
conv_final_layer = image_net['relu5_4']
pool5 = utils.max_pool_2x2(conv_final_layer)
pool5_flatten = tf.reshape(pool5, [-1, 7 * 7 * 512])
W1 = utils.get_weights_variable(7 * 7 * 512, 4096, name='W1')
b1 = utils.get_bias_variable(4096, name='b1')
matmul1 = tf.nn.relu(tf.nn.bias_add(tf.matmul(pool5_flatten, W1), b1))
matmul1 = tf.nn.dropout(matmul1, keep_prob)
W2 = utils.get_weights_variable(4096, 4096, name='W2')
b2 = utils.get_bias_variable(4096, name='b2')
matmul2 = tf.nn.bias_add(tf.matmul(matmul1, W2), b2)
matmul2 = tf.nn.dropout(matmul2, keep_prob)
W3 = utils.get_weights_variable(4096, CLASS_COUNTS, name='W3')
b3 = utils.get_bias_variable(CLASS_COUNTS, name='b3')
matmul3 = tf.nn.bias_add(tf.matmul(matmul2, W3), b3)
return matmul3
def classifier1(feature, keep_prob):
with tf.name_scope('classifier1'):
W1 = utils.get_weights_variable(4096, 1024, name='W1')
b1 = utils.get_bias_variable(1024, name='b1')
matmul1 = tf.nn.relu(tf.nn.bias_add(tf.matmul(feature, W1), b1))
matmul1 = tf.nn.dropout(matmul1, keep_prob)
W2 = utils.get_weights_variable(1024, CLASS_COUNTS, name='W2')
b2 = utils.get_bias_variable(CLASS_COUNTS, name='b2')
matmul2 = tf.nn.bias_add(tf.matmul(matmul1, W2), b2)
return matmul2
def Domain1(features, grad_scale, keep_prob):
with tf.name_scope('Domain1'):
features = flip_gradient(features, grad_scale)
W1 = utils.get_weights_variable(4096, 1024, name='W1')
b1 = utils.get_bias_variable(1024, name='b1')
matmul1 = tf.nn.relu(tf.nn.bias_add(tf.matmul(features, W1), b1))
matmul1 = tf.nn.dropout(matmul1, keep_prob)
W2 = utils.get_weights_variable(1024, 2, name='W2')
b2 = utils.get_bias_variable(2, name='b2')
matmul2 = tf.nn.bias_add(tf.matmul(matmul1, W2), b2)
return matmul2
def NetworkModel(image, keep_prob, reuse=False):
with tf.variable_scope('NetworkModel', reuse=reuse):
def _vgg_net(weights, image):
layers = (
'conv1_1', 'relu1_1', 'conv1_2', 'relu1_2', 'pool1',
'conv2_1', 'relu2_1', 'conv2_2', 'relu2_2', 'pool2',
'conv3_1', 'relu3_1', 'conv3_2', 'relu3_2', 'conv3_3',
'relu3_3', 'conv3_4', 'relu3_4', 'pool3',
'conv4_1', 'relu4_1', 'conv4_2', 'relu4_2', 'conv4_3',
'relu4_3', 'conv4_4', 'relu4_4', 'pool4',
'conv5_1', 'relu5_1', 'conv5_2', 'relu5_2', 'conv5_3',
'relu5_3', 'conv5_4', 'relu5_4'
)
net = {}
current = image
for i, name in enumerate(layers):
kind = name[:4]
if kind == 'conv':
kernels, bias = weights[i][0][0][0][0]
kernels = utils.get_variable(
np.transpose(kernels, (1, 0, 2, 3)), name=name + '_w')
bias = utils.get_variable(bias.reshape(-1), name=name + '_b')
current = utils.conv2d_basic(current, kernels, bias)
tf.add_to_collection("losses", tf.contrib.layers.l2_regularizer(0.0005)(kernels))
elif kind == 'relu':
current = tf.nn.relu(current, name=name)
elif kind == 'pool':
current = utils.max_pool_2x2(current)
net[name] = current
return net
print('setting up vgg model initialized params --> extractor')
model_data = utils.get_model_data("data", MODEL_URL)
weights = np.squeeze(model_data['layers'])
with tf.name_scope('inference'):
image_net = _vgg_net(weights, image)
conv_final_layer = image_net['relu5_4']
pool5 = utils.max_pool_2x2(conv_final_layer)
pool5_flatten = tf.reshape(pool5, [-1, 7 * 7 * 512])
W1 = utils.get_weights_variable(7 * 7 * 512, 4096, name='W1')
tf.add_to_collection("losses", tf.contrib.layers.l2_regularizer(0.0005)(W1))
b1 = utils.get_bias_variable(4096, name='b1')
matmul1 = tf.nn.bias_add(tf.matmul(pool5_flatten, W1), b1)
matmul1 = tf.nn.relu(matmul1)
matmul1 = tf.nn.dropout(matmul1, keep_prob)
W2 = utils.get_weights_variable(4096, 4096, name='W2')
tf.add_to_collection("losses", tf.contrib.layers.l2_regularizer(0.0005)(W2))
b2 = utils.get_bias_variable(4096, name='b2')
matmul2 = tf.nn.bias_add(tf.matmul(matmul1, W2), b2)
matmul2 = tf.nn.relu(matmul2)
matmul2 = tf.nn.dropout(matmul2, keep_prob)
W3 = utils.get_weights_variable(4096, CLASS_COUNTS, name='W3')
tf.add_to_collection("losses", tf.contrib.layers.l2_regularizer(0.0005)(W3))
b3 = utils.get_bias_variable(CLASS_COUNTS, name='b3')
pre_logits = tf.nn.bias_add(tf.matmul(matmul2, W3), b3)
# logits = tf.nn.softmax(pre_logits)
# regularizer_losses = tf.get_collection("losses")
return pre_logits #, logits
x_image = tf.reshape(x, [-1, img_size, img_size, num_channels])
y_true = tf.placeholder(tf.float32, shape=[None, num_classes], name='y_true')
y_true_cls = tf.argmax(y_true, dimension=1)
x_pretty = pt.wrap(x_image)
with pt.defaults_scope(activation_fn=tf.nn.relu):
y_pred, loss = x_pretty.\
conv2d(kernel=5, depth=16, name='layer_conv1').\
max_pool(kernel=2, stride=2).\
conv2d(kernel=5, depth=36, name='layer_conv2').\
max_pool(kernel=2, stride=2).\
flatten().\
fully_connected(size=128, name='layer_fc1').\
softmax_classifier(num_classes=num_classes, labels=y_true)
weights_conv1 = utils.get_weights_variable(layer_name='layer_conv1')
weights_conv2 = utils.get_weights_variable(layer_name='layer_conv2')
optimizer = tf.train.AdamOptimizer(learning_rate=1e-4).minimize(loss)
y_pred_cls = tf.argmax(y_pred, dimension=1)
correct_prediction = tf.equal(y_pred_cls, y_true_cls)
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
session = tf.Session()
session.run(tf.global_variables_initializer())
utils.print_test_accuracy(session=session,
x=x,
y_true=y_true,
y_pred_cls=y_pred_cls,
show_example_errors=False,