正文之前
在家闲着也是闲着，看论文的间隙，学习一下新鲜的深度学习框架也不错，参加几次学术会议，百度都在卖力的推广这个深度学习框架，其实最青睐的就是中文的文档啊，我可太爱了~
正文
没啥好说的， 从头看起，照着官方的文档实现了线性回归的房价预测，代码如下：
from __future__ import print_function
import paddle
import paddle.fluid as fluid
import numpy
import math
import sys
BATCH_SIZE = 20
train_reader = paddle.batch(
paddle.reader.shuffle(
paddle.dataset.uci_housing.train(), buf_size=500
),
batch_size=BATCH_SIZE
)
test_reader = paddle.batch(
paddle.reader.shuffle(
paddle.dataset.uci_housing.test(), buf_size=500
),
batch_size=BATCH_SIZE
)
x = fluid.layers.data(name='x', shape=[13], dtype='float32') # 定义输入的形状和数据类型
y = fluid.layers.data(name='y', shape=[1], dtype='float32') # 定义输出的形状和数据类型
y_predict = fluid.layers.fc(input=x, size=1) # 连接输入和输出的全连接层
main_program = fluid.default_main_program()
starup_program = fluid.default_startup_program()
# 利用标签数据和输出的预测数据估计方差
cost = fluid.layers.square_error_cost(input=y_predict, label=y)
# 对方差求均值，得到平均损失
avg_loss = fluid.layers.mean(cost)
test_program = main_program.clone(for_test=True)
sgd_optimizer = fluid.optimizer.SGD(learning_rate=0.001)
sgd_optimizer.minimize(avg_loss)
use_cuda = False
# 指明executor的执行场所
place = fluid.CUDAPlace(0) if use_cuda else fluid.CPUPlace()
###executor可以接受传入的program，
# 并根据feed map(输入映射表)和fetch list(结果获取表)
# 向program中添加数据输入算子和结果获取算子。
###使用close()关闭该executor，调用run(...)执行program。
exe = fluid.Executor(place)
num_epochs = 100
def train_test(executor, program, reader, feeder, fetch_list):
accumulated = 1 * [0]
count = 0
for data_test in reader():
outs = executor.run(program = program,
feed = feeder.feed(data_test),
fetch_list = fetch_list )
# 累加测试过程中的损失值
accumulated = [x_c[0] + x_c[1][0] for x_c in zip(accumulated,outs)]
# 累加测试过程中的样本数目
count += 1
return [x_d / count for x_d in accumulated]
%matplotlib inline
params_dirname = "fit_a_line.inference.model"
feeder = fluid.DataFeeder(place = place, feed_list=[x,y])
exe.run(starup_program)
train_prompt = "train cost"
test_prompt = "test cost"
from paddle.utils.plot import Ploter
plot_prompt = Ploter(train_prompt,test_prompt)
step = 0
exe_test = fluid.Executor(place)
for pass_id in range(num_epochs):
for data_train in train_reader():
avg_loss_value, = exe.run(
main_program,
feed = feeder.feed(data_train),
fetch_list=[avg_loss]
)
if step % 10 == 0:
plot_prompt.append(train_prompt, step, avg_loss_value[0])
plot_prompt.plot()
print("%s, Step %d, Cost %f" %
(train_prompt, step, avg_loss_value[0]))
if step % 100 == 0: # 每100批次记录并输出一下测试损失
test_metics = train_test(executor=exe_test,
program=test_program,
reader=test_reader,
fetch_list=[avg_loss.name],
feeder=feeder)
plot_prompt.append(test_prompt, step, test_metics[0])
plot_prompt.plot()
print("%s, Step %d, Cost %f" %
(test_prompt, step, test_metics[0]))
if test_metics[0] < 10.0: # 如果准确率达到要求，则停止训练
break
step += 1
if math.isnan(float(avg_loss_value[0])):
sys.exit("got NaN loss, training failed.")
#保存训练参数到之前给定的路径中
if params_dirname is not None:
fluid.io.save_inference_model(params_dirname, ['x'], [y_predict], exe)
train cost, Step 1480, Cost 30.048767
infer_exe = fluid.Executor(place)
inference_scope = fluid.core.Scope()
def save_result(points1, points2):
import matplotlib
matplotlib.use("Agg")
import matplotlib.pyplot as plt
x1 = [idx for idx in range(len(points1))]
y1 = points1
y2 = points2
l1 = plt.plot(x1, y1, 'r--', label='predictions')
l2 = plt.plot(x1, y2, 'g--', label='GT')
plt.plot(x1, y1, 'ro-', x1, y2, 'g+-')
plt.title('predictions VS GT')
plt.legend()
plt.savefig('image/prediction_gt.png')
with fluid.scope_guard(inference_scope):
[inference_program, feed_target_names,
fetch_targets] = fluid.io.load_inference_model(params_dirname, infer_exe)
#准备测试集
infer_reader = paddle.batch(
paddle.dataset.uci_housing.test(), batch_size = BATCH_SIZE)
infer_data = next(infer_reader())
print(infer_data[0][0])
print(infer_data[0][1])
infer_feat = numpy.array(
[data[0] for data in infer_data]).astype("float32") # 提取测试集中的数据
infer_label = numpy.array(
[data[1] for data in infer_data]).astype("float32") # 提取测试集中的标签
assert feed_target_names[0] == 'x'
results = infer_exe.run(inference_program,
feed={feed_target_names[0]: numpy.array(infer_feat)},
fetch_list=fetch_targets) # 进行预测
#打印预测结果和标签并且可视化结果：
print("infer results: (House Price)")
for idx, val in enumerate(results[0]):
print("%d: %.2f" % (idx, val)) # 打印预测结果
print("\nground truth:")
for idx, val in enumerate(infer_label):
print("%d: %.2f" % (idx, val)) # 打印标签值
save_result(results[0], infer_label) # 保存图片
[ 0.42616306 -0.11363636 0.25525005 -0.06916996 0.28457807 -0.14440207
0.17327599 -0.19893267 0.62828665 0.49191383 0.18558153 -0.0686218
0.40637243]
[8.5]
infer results: (House Price)
0: 13.95
1: 14.37
2: 13.85
3: 15.86
4: 14.36
5: 15.01
6: 14.24
7: 13.94
8: 11.17
9: 14.05
10: 10.63
11: 12.58
12: 13.36
13: 12.90
14: 12.77
15: 14.03
16: 15.73
17: 15.54
18: 15.77
19: 13.48
ground truth:
0: 8.50
1: 5.00
2: 11.90
3: 27.90
4: 17.20
5: 27.50
6: 15.00
7: 17.20
8: 17.90
9: 16.30
10: 7.00
11: 7.20
12: 7.50
13: 10.40
14: 8.80
15: 8.40
16: 16.70
17: 14.20
18: 20.80
19: 13.40
正文之后
好久没写过简书了，今天看到了那啥日更挑战，就来注水吧~
