Question

尝试导入numpy并调用np.polynomial.polynomial

import gym
import tensorflow as tf
import numpy as np
print(np.__version__)
methods =[method_name for method_name in dir(np.polynomial)
 if callable((getattr(np.polynomial, method_name)))]
print(methods)
polyvalfromroots = np.polynomial.polynomial.polyvalfromroots
from policy_gradient import util
from policy_gradient.policy import CategoricalPolicy

结果：

1.13.1
['Chebyshev', 'Hermite', 'HermiteE', 'Laguerre', 'Legendre', 'Polynomial', '_numpy_tester', 'bench', 'test']

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-26-41682dd8722b> in <module>()
      6  if callable((getattr(np.polynomial, method_name)))]
      7 print(methods)
----> 8 polyvalfromroots = np.polynomial.polynomial.polyvalfromroots
      9 from policy_gradient import util
     10 from policy_gradient.policy import CategoricalPolicy

AttributeError: module 'numpy.polynomial' has no attribute 'polynomial'

它已被修改为np.polynomial.Polynomial。

但文档here仍然写numpy.polynomial.polynomial

这是一个错误吗？我应该使用numpy.polynomial.polynomial恢复哪个版本的numpy？

Answer 1

我不知道为什么我会得到那个bug。但它通过重新安装我的numpy

解决了

conda remove numpy
conda install numpy=1.13

（Python）numpy.polynomial.polynomial是否在numpy 1.13.1版本中被破坏了？

1 个答案: