Follow these steps to start working with Jupyter Notebooks:
Activate the framework(s) you want to use
source /opt/DL/caffe-ibm/bin/caffe-activate # Caffe source /opt/DL/tensorflow/bin/tensorflow-activate # Tensorflow source /opt/DL/theano/bin/theano-activate # Theano source /opt/DL/torch/bin/torch-activate # Torch source /opt/DL/digits/bin/digits-activate # DIGITS source /opt/DL/nccl/bin/nccl-activate # For multi-GPU applications
Start a Jupyter Notebook server
jupyter-notebook --ip=0.0.0.0 --no-browser
You will see an output like this:
... [I 11:31:58.994 NotebookApp] The Jupyter Notebook is running at: [I 11:31:58.994 NotebookApp] http://0.0.0.0:8888/?token=e7d571edefbac3d82a91f2b1f8ee3250260ce3652e1e6b49 [I 11:31:58.994 NotebookApp] Use Control-C to stop this server and shut down all kernels (twice to skip confirmation). [C 11:31:58.995 NotebookApp] Copy/paste this URL into your browser when you connect for the first time, to login with a token: http://0.0.0.0:8888/?token=e7d571edefbac3d82a91f2b1f8ee3250260ce3652e1e6b49 ...
Copy the address into your browser, but change 0.0.0.0 to the address of the server you are using. For example:
# If you are using the Power 8 server http://p8.ncsa.illinois.edu:8888/?token=e7d571edefbac3d82a91f2b1f8ee3250260ce3652e1e6b49 # If you are using the Power 9 server http://hal000.ncsa.illinois.edu:8888/?token=e7d571edefbac3d82a91f2b1f8ee3250260ce3652e1e6b49
- You can now start working. If you want to put the Jupyter Notebook server in the background and regain control of the shell, do the following:
- Press Ctrl+Z in the terminal;
- Type bg and press enter to execute.
- Type fg in the terminal and execute. This will bring the server back to the foreground;
- Press Ctrl+C to shutdown the server;
- Answer y to the confirmation prompt.
Important Notes
- The Power 8 server currently only has Tensorflow 1.1.0. This is a limit of the current version of PowerAI, and may change in the future.