Due to the recent use of deep learning, it is necessary to install a Python-based deep learning framework. Python and some scientific computing packages (numpy, scipy) have been installed on the computer before, so I started work directly. I found a framework comparison about deep learning on the Internet. For details, please refer to this article: Contrasting the top ten frameworks of deep learning .
It can be seen that Keras and Caffe are now mainstream.
So enter the command on the command line:
pip install keras
Couldn't install due to network problems. So go to LFD to search for Keras wheels (.whl type files).
You can choose the version corresponding to your Python. This will successfully install.
Of course, if the installation reports an error, it is basically due to the lack of some dependent packages. At this time, you can search for the corresponding whl file (I have also installed TensorFlow and Theano before installing Keras, in case any dependencies are needed.)
Later, I went to the official website and looked it up, and it is indeed like this:
Then enter the Python environment to test whether the installation is successful.
Since it is built on the tensorflow framework by default, it showsUsing TensorFlow backend.