Keras is an advanced neural network API written in Python that can run with TensorFlow, CNTK, or Theano as a backend.Keras was developed with a focus on supporting fast experiments. Being able to convert your ideas into experimental results with minimal latency is key to good research. Install a Keras-compatible version of Python: Python 2.7-3.6.
I am using pycharm-2019.3 version
Step one. Directly under the command windowpip install keras==2.1.6
Here I have set up Ali's mirror, if you use other mirrors, or did not set up a mirror before, you can use the commandpip instll -i /simple keras==2.1.6Of course, you can also import it from the settings
Version 2.1.6 will do.
That's what happened after my run here.
Installation from GitHub also mention it, but really do not recommend - slow. And the western capital of that urine, not here to complain more, we can later develop a good code cloud.
First, use git to clone Keras:
git clone /keras-team/
Then, cd to the Keras directory and run the install command:
cd keras sudo python install
Step two. Configure the Keras backend
On windows, in your personal folder (C:/Users/%your username% if you haven't changed the default windows settings).keras\
Use of tensorflow and theano in keras-based model training
If keras is used to build the model, you need to configure ~/.keras/ and keras needs to specify the backend as follows:
{
“image_dim_ordering”: “tf”,
“epsilon”: 1e-07,
“floatx”: “float32”,
“backend”: “tensorflow”
}
This configuration uses tensorflow as the backend
With theano as the backend, the configuration is as follows:
{
“image_dim_ordering”: “th”,
“epsilon”: 1e-07,
“floatx”: “float32”,
“backend”: “theano”
}
For more information, see the keras documentation.keras documentation
To install tensorflow, here's my personal recommendation.
Recommended python version 3.6
I was on 3.7 and couldn't get it to work. 3.8, bye bye!
pip install -i /simple tensorflow==1.8.0
Don't ask why is 1.8 version, ask is limited energy various versions are not compatible with the problem, this is the highest version of our side successfully installed.
pip install -i /simple numpy==1.16.0
Don't ask why version 1.16, ask is limited energy various versions are not compatible with the problem, this is the highest version of our side successfully installed. If you have any other version, you can uninstall it.
To this point this article on the analysis of the installation of Keras (pycharm) and the initial understanding of the article is introduced to this, more related to the installation of Keras content, please search for my previous articles or continue to browse the following related articles I hope you will support me in the future more!