Deep Learning Day 1 – Install Tensorflow, Keras and Jupyter for dev

This post is focussed on getting our dev env setup for our future  experiments.

Dev stack is pretty minimal for now:

  • TensorFlow
  • Keras
  • Jupyter notebooks

Get virtual env setup

$> virtualenv -p python3 venv

Installing dependencies

These are the dependencies we would use soon as part of our play tasks.

$> vim requirements.txt

numpy
scipy
tensorflow==0.12.0
tensorflow-gpu
keras
ipython
jupyter
six
protobuf==3.1.0
wheel

$> pip install -r requirements.txt

Run Jupyter notebook

$> jupyter notebook

Screen Shot 2017-12-20 at 11.53.24 PM

 

Screen Shot 2017-12-20 at 11.53.09 PM

Try this in your Jupyter notebook:

 

Screen Shot 2017-12-21 at 9.24.17 AM 1

 

More (troubleshooting):

Sometimes import tensorflow still fails. You can try installing dependencies like:

-- import tensorflow as tf error:
export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.12.0-py3-none-any.whl
sudo pip3 install --upgrade $TF_BINARY_URL

Leave a Reply

Your email address will not be published. Required fields are marked *