paddle 基础函数 cross_entropy

官方文档：https://www.paddlepaddle.org.cn/documentation/docs/zh/api_cn/layers_cn/cross_entropy_cn.html

 

示例：

import paddle.fluid as fluid
import numpy as np

class_num = 7
x = fluid.data(name='x', shape=[-1, 1, 10], dtype='float32')
label = fluid.data(name='label', shape=[-1, 1], dtype='int64')
predict = fluid.layers.fc(input=x, size=class_num, act='softmax')
cost = fluid.layers.cross_entropy(input=predict, label=label)

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

samples = 6
np_x = np.random.random(size=(samples, 1, 10)).astype('float32')
np_y = np.random.randint(low=0, high=class_num, size=(samples, 1)).astype('int64')
print(np_x)
print(np_y)
output = exe.run(feed={"x": np_x, "label": np_y}, fetch_list = [cost])
print(output)

结果：

[[[0.4527526  0.79921854 0.4485681  0.34811467 0.7507561  0.8078209   0.72739184 0.5410976  0.4951061  0.08417795]]

 [[0.69459516 0.5052644  0.09735702 0.17990156 0.04351452 0.4255496   0.34425682 0.5052834  0.9365378  0.06064539]]

 [[0.09274481 0.9413779  0.3883635  0.65820855 0.60970956 0.13670038  0.64331406 0.08975768 0.7772037  0.8579608 ]]

 [[0.68027896 0.90740174 0.02467881 0.94697666 0.84760493 0.90569997  0.4304838  0.43157798 0.44898176 0.77926594]]

 [[0.94457364 0.27906907 0.913569   0.5994569  0.1048044  0.28563756  0.3188055  0.68807054 0.6438866  0.01220882]]

 [[0.24974725 0.2163003  0.16348869 0.8209394  0.671093   0.30101737  0.3471077  0.79650295 0.02727815 0.40069225]]]


[[4]
 [0]
 [6]
 [0]
 [4]
 [2]]


[array([[1.5678651],
       [1.8718125],
       [2.6164384],
       [1.7536728],
       [1.7999921],
       [1.4771321]], dtype=float32)]