def main():
paddle.init(use_gpu=False, trainer_count=1)
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear())
y = paddle.layer.data(name='y', type=paddle.data_type.dense_vector(1))
cost = paddle.layer.mse_cost(input=y_predict, label=y)
parameters = paddle.parameters.create(cost)
optimizer = paddle.optimizer.Momentum(momentum=0)
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer)
feeding = {'x': 0, 'y': 1}
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 100 == 0:
print "Pass %d, Batch %d, Cost %f" % (event.pass_id, event.batch_id,
event.cost)
if isinstance(event, paddle.event.EndPass):
result = trainer.test(reader=paddle.batch(uci_housing.test(),
batch_size=2),
feeding=feeding)
print "Test %d, Cost %f" % (event.pass_id, result.cost)
trainer.train(reader=paddle.batch(
paddle.reader.shuffle(uci_housing.train(),
buf_size=500),
batch_size=2),
feeding=feeding,
event_handler=event_handler,
num_passes=30)
def main():
# init
paddle.init(use_gpu=False, trainer_count=1)
# network config
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x,
param_attr=paddle.attr.Param(name='w'),
size=1,
act=paddle.activation.Linear(),
bias_attr=paddle.attr.Param(name='b'))
y = paddle.layer.data(name='y', type=paddle.data_type.dense_vector(1))
cost = paddle.layer.mse_cost(input=y_predict, label=y)
# create parameters
parameters = paddle.parameters.create(cost)
# create optimizer
optimizer = paddle.optimizer.Momentum(momentum=0)
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer)
# event_handler to print training and testing info
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 100 == 0:
print "Pass %d, Batch %d, Cost %f" % (
event.pass_id, event.batch_id, event.cost)
if isinstance(event, paddle.event.EndPass):
if (event.pass_id + 1) % 10 == 0:
result = trainer.test(reader=paddle.batch(uci_housing.test(),
batch_size=2),
feeding={
'x': 0,
'y': 1
})
print "Test %d, %.2f" % (event.pass_id, result.cost)
# training
trainer.train(reader=paddle.batch(paddle.reader.shuffle(
uci_housing.train(), buf_size=500),
batch_size=2),
feeding={
'x': 0,
'y': 1
},
event_handler=event_handler,
num_passes=30)
def main():
# init
paddle.init(use_gpu=False, trainer_count=1)
# network config
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x,
param_attr=paddle.attr.Param(name='w'),
size=1,
act=paddle.activation.Linear(),
bias_attr=paddle.attr.Param(name='b'))
y = paddle.layer.data(name='y', type=paddle.data_type.dense_vector(1))
cost = paddle.layer.mse_cost(input=y_predict, label=y)
# create parameters
parameters = paddle.parameters.create(cost)
# create optimizer
optimizer = paddle.optimizer.Momentum(momentum=0)
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer)
# event_handler to print training and testing info
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 100 == 0:
print "Pass %d, Batch %d, Cost %f" % (
event.pass_id, event.batch_id, event.cost)
if isinstance(event, paddle.event.EndPass):
if (event.pass_id + 1) % 10 == 0:
result = trainer.test(
reader=paddle.batch(
uci_housing.test(), batch_size=2),
feeding={'x': 0,
'y': 1})
print "Test %d, %.2f" % (event.pass_id, result.cost)
# training
trainer.train(
reader=paddle.batch(
paddle.reader.shuffle(
uci_housing.train(), buf_size=500),
batch_size=2),
feeding={'x': 0,
'y': 1},
event_handler=event_handler,
num_passes=30)
def main():
# init
paddle.init(use_gpu=with_gpu, trainer_count=1)
# network config
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13))
y_predict = paddle.layer.fc(input=x,
size=1,
act=paddle.activation.Linear())
y = paddle.layer.data(name='y', type=paddle.data_type.dense_vector(1))
cost = paddle.layer.square_error_cost(input=y_predict, label=y)
# Save the inference topology to protobuf.
inference_topology = paddle.topology.Topology(layers=y_predict)
with open("inference_topology.pkl", 'wb') as f:
inference_topology.serialize_for_inference(f)
# create parameters
parameters = paddle.parameters.create(cost)
# create optimizer
optimizer = paddle.optimizer.Momentum(momentum=0)
trainer = paddle.trainer.SGD(cost=cost,
parameters=parameters,
update_equation=optimizer)
feeding = {'x': 0, 'y': 1}
# event_handler to print training and testing info
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 100 == 0:
print "Pass %d, Batch %d, Cost %f" % (
event.pass_id, event.batch_id, event.cost)
if isinstance(event, paddle.event.EndPass):
if event.pass_id % 10 == 0:
with open('params_pass_%d.tar' % event.pass_id, 'w') as f:
trainer.save_parameter_to_tar(f)
result = trainer.test(reader=paddle.batch(uci_housing.test(),
batch_size=2),
feeding=feeding)
print "Test %d, Cost %f" % (event.pass_id, result.cost)
# training
trainer.train(reader=paddle.batch(paddle.reader.shuffle(
uci_housing.train(), buf_size=500),
batch_size=2),
feeding=feeding,
event_handler=event_handler,
num_passes=30)
# inference
test_data_creator = paddle.dataset.uci_housing.test()
test_data = []
test_label = []
for item in test_data_creator():
test_data.append((item[0], ))
test_label.append(item[1])
if len(test_data) == 5:
break
# load parameters from tar file.
# users can remove the comments and change the model name
# with open('params_pass_20.tar', 'r') as f:
# parameters = paddle.parameters.Parameters.from_tar(f)
probs = paddle.infer(output_layer=y_predict,
parameters=parameters,
input=test_data)
for i in xrange(len(probs)):
print "label=" + str(test_label[i][0]) + ", predict=" + str(
probs[i][0])
def main():
# 0.init paddle初始化定义跑模型的设备
paddle.init(use_gpu=with_gpu, trainer_count=1)
# 1.读取data数据
x = paddle.layer.data(name='x', type=paddle.data_type.dense_vector(13)) #13列feature数据层, dense_vector：稠密的浮点数向量
y = paddle.layer.data(name='y', type=paddle.data_type.dense_vector(1)) #1列label数据层, dense_vector：稠密的浮点数向量
# 2.network config 定义网络结构（创建一个全连接层，激活函数为线性函数）
y_predict = paddle.layer.fc(input=x, size=1, act=paddle.activation.Linear())
# 3.定义cost损失函数，此处使用square平方差cost函数
cost = paddle.layer.square_error_cost(input=y_predict, label=y)
# 4.保存网络拓扑到protobuf
inference_topology = paddle.topology.Topology(layers=y_predict)
with open("inference_topology.pkl", 'wb') as f:
inference_topology.serialize_for_inference(f)
# 5.create parameters指定训练相关的参数
parameters = paddle.parameters.create(cost)
# 6.create optimizer定义训练方法(Momentum：梯度下降方向与上次一致时加大下降幅度，已加速收敛)
optimizer = paddle.optimizer.Momentum(momentum=0)
# 7.指定trainer训练优化器(SGD:随机梯度下降stochastic gradient descent)
trainer = paddle.trainer.SGD(
cost=cost, parameters=parameters, update_equation=optimizer)
feeding = {'x': 0, 'y': 1}
# 8.定义event_handler，输出训练过程中的结果
lists = []
def event_handler(event):
if isinstance(event, paddle.event.EndIteration):
if event.batch_id % 100 == 0:
print "Pass %d, Batch %d, Cost %f" % (
event.pass_id, event.batch_id, event.cost)
if isinstance(event, paddle.event.EndPass):
#保存参数
#if event.pass_id % 10 == 0:
with open('params_pass_%d.tar' % event.pass_id, 'w') as f:
trainer.save_parameter_to_tar(f)
result = trainer.test(
reader=paddle.batch(uci_housing.test(), batch_size=2),
feeding=feeding)
print "Test %d, Cost %f" % (event.pass_id, result.cost)
#print result.metrics
#保存训练结果损失情况
lists.append((event.pass_id, result.cost,
#result.metrics['classification_error_evaluator']))
result.metrics))
# 9.training开始训练模型
#创建一个reader，实质上是一个迭代器，每次返回一条数据
reader = uci_housing.train()
#创建一个shuffle_reader，把上一步的reader放进去，配置buf_size就可以读取buf_size大小的数据自动做shuffle，让数据打乱，随机化
shuffle_reader = paddle.reader.shuffle(reader, buf_size=500)
#创建一个batch_reader，把shuffle后的数据，一个batch一个batch的形式，批量的放到训练器里去进行每一步的迭代和训练
batch_reader = paddle.batch(shuffle_reader, batch_size=2)
#开始训练
trainer.train(
#reader=paddle.batch(
# paddle.reader.shuffle(uci_housing.train(), buf_size=500),
# batch_size=2),
reader=batch_reader,
#reader=paddle.batch(reader, batch_size=2),
feeding=feeding,
event_handler=event_handler,
num_passes=30)
#找到训练误差最小的一次结果
best = sorted(lists, key=lambda list: float(list[1]))[0]
best_passid = best[0] #训练误差最小的模型passid
best_cost = best[1] #训练误差最小的损失函数返回cost
print 'Best pass is %s, testing Avgcost is %s' % (best_passid, best_cost)
#print str(best[2])
#print 'The classification accuracy is %.2f%%' % (100 - float(best[2]) * 100)
#清理非best模型参数文件
for tmp in lists:
if tmp[0] != best_passid:
os.remove('params_pass_%s.tar' % tmp[0])
# 10.inference使用模型进行预测
#加载测试数据
test_data_creator = paddle.dataset.uci_housing.test()
test_data = []
test_label = []
for item in test_data_creator():
test_data.append((item[0], ))
test_label.append(item[1])
#print item[1]
#if len(test_data) == 5:
# break
# load parameters from tar file.
# users can remove the comments and change the model name
# 使用训练误差最小那次的模型
#print parameters.__dict__
with open('params_pass_%s.tar' % best_passid, 'r') as f:
print 'Use best pass params: %s' % best_passid
parameters = paddle.parameters.Parameters.from_tar(f)
#print parameters.__dict__
#运行预测
probs = paddle.infer(
output_layer=y_predict, parameters=parameters, input=test_data)
#模型预测结果打印
for i in xrange(len(probs)):
#print "label=" + str(test_label[i][0]) + ", predicted price=${:,.2f}" . format(probs[i][0] * 1000)
print "label=" + str(test_label[i][0]) + ", predicted price=$%s" % probs[i][0]
plot_cost.append(test_title, step, result.cost)
if step % 100 == 0:
plot_cost.plot()
step += 1
if isinstance(event, paddle.event.EndPass):
if event.pass_id % 10 == 0:
with open('params_pass_%d.tar' % event.pass_id, 'w') as f:
parameters.to_tar(f)
# Starting the training process
trainer.train(reader=paddle.batch(paddle.reader.shuffle(uci_housing.train(),
buf_size=500),
batch_size=2),
feeding=feeding,
event_handler=event_handler_plot,
num_passes=30)
# generating testing data
test_data_creator = paddle.dataset.uci_housing.test()
test_data = []
test_label = []
for item in test_data_creator():
test_data.append((item[0], ))
test_label.append(item[1])
if len(test_data) == 5: #wtf?