Machine learning TDT4173

A quick demo of the Recurrent Neural Network-model using LSTM is available at https://tdt4173-project-sl-1.web.app/, alongside a decription and information about all the methods used in this project.

Running the models

Each of the three approaches have their own .ipynb-file. To run these, simply clone the project, add the data located in /data/ and update the file paths in the second code-block of each notebook, variables train and valid, to the correct path. The notebooks should now be ready to run sequentially.

ml-demo

The folder /ml-demo/ contains the code used to present our models in the above-mentioned website.

Running ml-demo

Download npm

https://www.npmjs.com/get-npm

Run the code

Clone the repository, navigate into the /ml-demo/-folder and run npm install. After installing dependencies, run npm start.

The app will now be running at http://localhost:3000. Unfortunately, the prediction do not work from localhost due to issues with CORS. We did not have the time to fix this.