- August 2017 - Identify a faculty advisor and a feasible project with an appropriate scope.
- September 2017 - Complete the first presentation (roughly project requirements).
- October 2017 - Complete second presentation (roughly revised requirements and design).
- November 2017 - Project implementation and testing, schedule final public presentation.
- December 2017 - Complete final presentation and upload all documents (design artifacts) and code to course file drop box for the Project Artifacts assignment.
September
- Pull a current subset of active decks
- Begin labeling labeling
- Begin initial design on NN
- Layers, convolution (?), output function
30 nodes per layer, 3 layers, activation=reLU, output=softmax
https://keras.io/getting-started/sequential-model-guide/ - Set up Keras, Tensor Flow, and Flask
October
- Begin front end work
- Flask serving static website
- Serve some JSON meta data
- Tinker with NN--layers, parameters
November
-
Link back end to NN for getting up-to-date infoNote: Unable to actually do this because HearthSim doesn't have even a non-public API that I can pull unlabled decks from. - Polish front end design
December
- Clean up work
- Reach out to HearthSim to see if they’re interested in this project
NN Docker Image:
https://github.com/floydhub/dl-docker
docker run -it -p 8888:8888 -p 6006:6006 -v ~/sharedfolder:/root/sharedfolder floydhub/dl-docker:cpu bash
jupyter notebook --allow-root
Flask:
http://flask.pocoo.org/docs/0.12/quickstart/#a-minimal-application