/tensorflow-intro

My Jupyter notebook from the Monash DataFluency Tensor Flow and Machine Learning workshop.

Primary LanguageJupyter Notebook

tensorflow-intro

My Jupyter notebook from the Monash DataFluency Tensor Flow and Machine Learning workshop.

The notebook makes use of the DataFluency TutorialSupport package and sample data.

To get started, clone this repository:

$ $ git clone https://github.com/cshaun/tensorflow-intro.git tensorflow-intro.git
$ cd ./tensorflow-intro.git

Then create a virtual environment and setup your environment:

$ python3 -m virtualenv --no-site-packages --distribute env36
$ source env/bin/activate
(env36) $ pip3 install -U git+https://github.com/MonashDataFluency/intro-to-tensorflow.git
(env36) $ pip3 install jupyter

This installs the Monash DataFluency TutorialSupport package and dependencies, including TensorFlow and some sample datasets for machine learning. Note that the TutorialSupport package resets the random seed each time a session is created, to ensure results are reproducible, and limits the number of threads to 2. The TutorialSupport package requires python3 (tested using python3.6 on OS X).

To start the Jupyter notebook server:

(env) $ jupyter notebook

which should launch your browser, displaying the Jupyter dashboard and the contents of the current directory.

Alternatively, you can probably open the notebook on Google's Colaboratory by clicking the button below:

Open In Colab.

To run on Colab you need to install the TutorialSupport package from within the notebook by adding and executing a code cell containing something like:

pip install git+https://github.com/MonashDataFluency/intro-to-tensorflow.git