无法在Google数据实验室中运行pmdarima。其错误的.ufunc大小已更改，

Question

我正在尝试从pmdarima.arima导入auto_arima运行。它给我一个错误numpy.ufunc大小更改，可能表明二进制不兼容。预期从C标头获得216，从PyObject获得192

安装pmdarima！pip安装pmdarima

   Requirement already up-to-date: pmdarima in /usr/local/envs/py3env/lib/python3.5/site-packages (1.6.0)
    Requirement already satisfied, skipping upgrade: numpy>=1.17.3 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pmdarima) (1.18.4)
    Requirement already satisfied, skipping upgrade: pandas>=0.19 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pmdarima) (0.22.0)
    Requirement already satisfied, skipping upgrade: urllib3 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pmdarima) (1.22)
    Requirement already satisfied, skipping upgrade: scikit-learn>=0.22 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pmdarima) (0.22.2.post1)
    Requirement already satisfied, skipping upgrade: scipy>=1.3.2 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pmdarima) (1.4.1)
    Requirement already satisfied, skipping upgrade: Cython>=0.29 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pmdarima) (0.29.17)
    Requirement already satisfied, skipping upgrade: statsmodels>=0.10.2 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pmdarima) (0.11.1)
    Requirement already satisfied, skipping upgrade: joblib>=0.11 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pmdarima) (0.14.1)
    Requirement already satisfied, skipping upgrade: python-dateutil>=2 in /usr/local/envs/py3env/lib/python3.5/site-packages (from pandas>=0.19->pmdarima) (2.5.0)
    Requirement already satisfied, skipping upgrade: pytz>=2011k in /usr/local/envs/py3env/lib/python3.5/site-packages (from pandas>=0.19->pmdarima) (2018.4)
    Requirement already satisfied, skipping upgrade: patsy>=0.5 in /usr/local/envs/py3env/lib/python3.5/site-packages (from statsmodels>=0.10.2->pmdarima) (0.5.0)
    Requirement already satisfied, skipping upgrade: six>=1.5 in /usr/local/envs/py3env/lib/python3.5/site-packages (from python-dateutil>=2->pandas>=0.19->pmdarima) (1.10.0)

numpy也被更新。

Answer 1

似乎是issue with numpy。您需要降级到1.16.1版本。

!pip install numpy==1.16.1

然后您需要重新启动内核，以使更改生效

import os
os._exit(00)