model-repo
frontend

Flask-backend for our Keras/Tensorflow model that predicts mushrooms based on pictures.

Running instance

Params: .jpg file as multipart/form-data

Returns: JSON-object that has prediction value on its edibility 0 being unedible and 1 edible.

Example:

{
  "prediction": 0.658562958240509
}

1. First install Python >=3.5
2. You should have now pip so then install virtualenv: pip install virtualenv (or pip3 install virtualenv if you have also Python 2 installed eg. Mac OS X)
3. Create virtualenv environment: virtualenv <name>
4. Activate that environment: source ./<name>/bin/activate
5. Clone this repository including the frontend with: git clone --recursive https://github.com/TeemuKoivisto/deep-shrooms-backend
6. Install requirements: pip install -r requirements.txt
7. Follow the installation instructions on the deep-shrooms-frontend
8. And you're done!

Run the development server on localhost:9000 with ./dev.sh.
It will automatically refresh on any changes to the source code.

Production server uses pm2 for running the app as a daemon. You can install it with npm i -g pm2.

To run the production server on localhost:9000 with pm2 start ./prod.sh.
NOTE: You have to build the frontend by going to its folder and entering: npm run build.

You can probably install the same requirements as in requirements.txt with:

pip install keras tensorflow flask flask-cors h5py gunicorn