def CreateModel(model_name, bit, use_gpu):
if model_name == 'vgg11':
vgg11 = models.vgg11(pretrained=True)
cnn_model = CNN_model.cnn_model(vgg11, model_name, bit)
if model_name == 'alexnet':
alexnet = models.alexnet(pretrained=True)
cnn_model = CNN_model.cnn_model(alexnet, model_name, bit)
if use_gpu:
cnn_model = cnn_model.cuda()
return cnn_model
def get_cnn_model(model_name, bits):
if model_name == 'vgg11':
vgg11 = models.vgg11(pretrained=True)
cnn_model = CNN_model.cnn_model(vgg11, model_name, bits)
if model_name == 'alexnet':
alexnet = models.alexnet(pretrained=True)
cnn_model = CNN_model.cnn_model(alexnet, model_name, bits)
if model_name == 'resnet':
cnn_model = CNN_model.getResnetModel(bits)
if torch.cuda.is_available():
cnn_model = cnn_model.cuda()
return cnn_model
def CreateModel(model_name, bit, use_gpu):
if model_name == 'vgg11':
vgg11 = models.vgg16(pretrained=True)
cnn_model = CNN_model.cnn_model(vgg11, model_name, bit)
if model_name == 'alexnet':
alexnet = models.alexnet(pretrained=True)
cnn_model = CNN_model.cnn_model(alexnet, model_name, bit)
if model_name == 'resnet34':
resnet = models.resnet34(pretrained=True)
cnn_model = CNN_model.cnn_model(resnet, model_name, bit)
if model_name == 'resnet50':
resnet = models.resnet50(pretrained=True)
cnn_model = CNN_model.cnn_model(resnet, model_name, bit)
if model_name == 'resnet34resnet34':
resnet = models.resnet34(pretrained=True)
cnn_model = CNN_model.cnn_model(resnet, model_name, bit)
return cnn_model
def CreateModel(model_name, bit, use_gpu):
if model_name == 'vgg16':
vgg16 = models.vgg16(pretrained=True)
#vgg16 = torch.load('/home/zhangjingyi/Rescode/vgg16_caffe2pytorch/vgg16_20M.pkl')
cnn_model = CNN_model.cnn_model(vgg16, model_name, bit)
print('**********************************')
print('Def of poo4~poo5:')
for i in [23, 24, 26, 28, 30]:
print(cnn_model.features[i])
print('**********************************')
if use_gpu:
cnn_model = torch.nn.DataParallel(cnn_model).cuda()
#cnn_model = cnn_model.cuda()
return cnn_model
def train_net(initial_learning_rate, loss_mode, optimizer_mode, data):
tf.reset_default_graph()
train_data, test_desired_input, test_desired_output = data
desired_input, desired_output, layer9, keep_prob = CNN_model.cnn_model(
).cnn_onemap_model()
model_name = "cnn_onemap_model"
global_ = tf.Variable(tf.constant(0))
decay_learning_rate = tf.train.exponential_decay(initial_learning_rate,
global_,
5,
0.95,
staircase=True)
loss, loss_name = loss_function(loss_mode, desired_output, layer9)
tf.summary.scalar("loss", loss)
train_step = optimizer(optimizer_mode,
decay_learning_rate)[0].minimize(loss)
optimizer_name = optimizer(optimizer_mode, decay_learning_rate)[1]
correct_prediction = tf.equal(tf.arg_max(desired_output, 1),
tf.arg_max(layer9, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
tf.summary.scalar("accuracy", accuracy)
#tensorboard merged all
merged = tf.summary.merge_all()
sess = tf.Session()
init = tf.global_variables_initializer()
sess.run(init)
writer = tf.summary.FileWriter(
LOGDIR + model_name + '_' + str(initial_learning_rate) + '_' +
loss_name + '_' + optimizer_name + "/", sess.graph)
writer.add_graph(sess.graph)
data_file = open(
LOGDIR + '/' + str(initial_learning_rate) + '_' + loss_name + '_' +
optimizer_name + ".txt", 'w')
for i in range(2):
minibatch = random.sample(train_data, 100)
image_batch = [d[0] for d in minibatch]
label_batch = [d[1] for d in minibatch]
sess.run(train_step,
feed_dict={
desired_input: image_batch,
desired_output: label_batch,
keep_prob: 0.5
})
sess.run(decay_learning_rate, feed_dict={global_: i})
training_loss = sess.run(loss,
feed_dict={
desired_input: image_batch,
desired_output: label_batch,
keep_prob: 1
})
test_loss = sess.run(loss,
feed_dict={
desired_input: test_desired_input,
desired_output: test_desired_output,
keep_prob: 1.0
})
accuracy_num = sess.run(accuracy,
feed_dict={
desired_input: test_desired_input,
desired_output: test_desired_output,
keep_prob: 1.0
})
#accuracy_num = sess.run(accuracy,feed_dict={desired_input:image_batch,take_desired_output:label_batch,keep_prob: 1.0})
print("step:", i, "accuracy:", accuracy_num, "training_loss",
training_loss, "test_loss", test_loss)
data_file.write(
str(i) + ',' + str(accuracy_num) + ',' + str(training_loss) + ',' +
str(test_loss) + '\n')
summary_data = sess.run(merged,
feed_dict={
desired_input: image_batch,
desired_output: label_batch,
keep_prob: 1
})
writer.add_summary(summary_data, i)