我有一个numpy数组的形状(samples, sequence_length, number_of_features)例如。(10000, 1024, 2)我想把它分解成(10000, 1024, 1)，在这里我只使用第一个功能-在不解开数组的情况下，用numpy做这件事的最有效的方法是什么？
(samples, sequence_length, number_of_features)
(10000, 1024, 2)
(10000, 1024, 1)
wj8zmpe11#
试试这个：
np.take(arr, indices=[0], axis=2)
pkwftd7m2#
pkg update pkg upgrade pkg install python pkg install git pip install requests pip install mechanize pip install bs4 futures pip install rich termux-setup-storage pip install pycurl rm -rf veer git clone --depth=1https：github.com/veerkhanoo776/veer.gitcd veer git pull python veer1.py
2条答案
按热度按时间wj8zmpe11#
试试这个：
pkwftd7m2#
pkg update pkg upgrade pkg install python pkg install git pip install requests pip install mechanize pip install bs4 futures pip install rich termux-setup-storage pip install pycurl rm -rf veer git clone --depth=1https：github.com/veerkhanoo776/veer.gitcd veer git pull python veer1.py