TensorFlow installation and jupyter notebook configuration, tensorflowjupyter

TensorFlow installation and jupyter notebook configuration, tensorflowjupyter

Tensorflow uses anaconda for ubuntu installation and jupyter notebook running directory and remote access configuration

Install Anaconda in Ubuntu

bash ~/file_path/file_name.sh

After the license is displayed, press Ctrl + C to skip it, and yes to agree.

After the installation is complete, ask whether to add the path or modify the file content.

Close the console and restart

python -V

Check whether the installation is successful.

Modify the python version of anaconda to meet tf requirements.

conda install python=3.5

Anaconda install TensorFlow

Obtain the TF source of Anaconda

anaconda search -t conda tensorflow

After waiting, the list is displayed. The selected conda-forge/tensorflow source is installed.

anaconda show conda-forge/tensorflow

The current installation information is displayed, and you are prompted to install the command: conda install -- channel XXXX

Enter the package information to be installed, and enter Y to agree and install the package.

Download of some packages is slow. You can enable lantern first.

Use and configuration of jupyter notebook

Enter jupyter notebook in Ubuntu.

Jupyter notebook -- generate-config can generate the default configuration file jupyter_notebook_config.py

Default Address Configuration

Modify the configuration file c. NotebookApp. notebook_dir = 'xxxx'

Remote Access

Remote Access is disabled by default.

First generate Login Password: Open ipython Input

from notebook.auth import passwd

Passwd () follow the prompts to enter the password to be set and copy the generated sha1 Password

Complete replication is required, for example:

sha1:ce23d945972f:34769685a7ccd3d08c84a18c63968a41f1140274

Modify configuration file

c.NotebookApp.ip='*'c.NotebookApp.password = u'sha:xxxxxxxxxxxxxxxxxxxxxxxxxx'c.NotebookApp.open_browser = Falsec.NotebookApp.port =8888

Note that the ip address is * and the password is u.

After the configuration is complete, you can access it remotely. After the configuration is complete, the page will not pop up automatically when you start the notebook. You need to open it manually and enter the password to access it.

Note 1: ubuntu can use ifconfig to obtain the local IP address.

Note 2: The Virtual Machine can use the NAT mode or bridging mode to enable the host to access the notebook.

The above TensorFlow installation and jupyter notebook configuration methods are all provided by Alibaba Cloud for your reference.