To run this project, either use a locally installed jupyter with tensorflow 2, keras, sklearn, numpy and pygraphviz, or simply run docker-compose up in this directory and use the provided docker image. It will have everything you need.
If you're new to Jupyter notebooks, running this command will create a running notebook server on your machine, exposed at port 8888.
After running this command, you can go to http://127.0.0.1:8888 to see your server.
You'll notice it will require a token for you to access it - you can find that token in the output logs of the docker-compose command. It will look something like this:
http://127.0.0.1:8888/?token=ecf605284aee8faf1e7edd6b647a7ba5d2bc3b539938e07e
You can in some terminals also just click the link to open it and automatically log in.
The exercises are labelled with numbers and I recommend starting with the first.
If you're new to how jupyter notebooks work, I recommend the interactive tutorial at: https://jupyter.org/try (pick the classic one). It is a bit lengthy, and covers most features. We'll only need a subset, so you can also just experiment here.
Have fun!
If you're looking to install this manually on your machine you'll need:
- Python 3.6 or 3.7 (3.8 not supported yet by tensorflow 2)
- Jupyter
- Graphviz installed (For windows, remember to add it to your PATH environment variable)
- pip3
For windows, go here: https://graphviz.gitlab.io/_pages/Download/Download_windows.html For mac: You can install graphviz using homebrew For ubuntu: Just apt-get install graphviz
- pygraphviz
- numpy
- tensorflow==2.1.0
- sklearn
- matplotlib