Created by: mapingshuo

复现代码：

import paddle
import paddle.fluid as fluid

x = fluid.layers.data(name='x', shape=[5, 4])
fc = fluid.layers.fc(input=x, size=7, bias_attr=None)
loss = fluid.layers.mean(fc)

sgd = fluid.optimizer.SGD(learning_rate = 0.01)
sgd.minimize(loss)

place = fluid.CPUPlace()
exe = fluid.Executor(place)

exe.run(fluid.default_startup_program())

fluid.io.save_params(exe,
                     "demo_model_dir")

with open("main_program", "w") as f:
  f.write(str(fluid.default_main_program()))

而文档中说：“如果 bias_attr 非空，则会创建一个偏置变量（bias variable），并把它累加到输出结果中。”