danielgchen / dubhacks2021_toneit

Quick Start

This is to just download the ToneIt app and utilize the example

1. Clone the repository
2. Download (the model)[https://drive.google.com/file/d/1IpBXq09vFqbJ2RRlDDlFCmHbU3WlnpCf/view?usp=sharing] (50epochs, 16000objects)
3. Run the main application via the GoogleCOLAB through the MainApp.ipynb file

Extensive Start

This is to build the model from scratch yourself

1. Clone the repository
2. Open the CreateModel.ipynb file in a GoogleCOLAB environment
3. Upload the trainTone.txt file to the GoogleCOLAB environment
4. Run the CreateModel.ipynb file (you can change parameters based on the HyperParameterTuning.ipynb file)
5. Download the ToneItPipeline.tar.gz archive from the GoogleCOLAB environment
6. Proceed to QuickStart with your ToneItPipeline model

Notes

• Model evalution metrics are found in CreateModel.ipynb file for the 50epochs 16000objects model
• More performance and evaluation metrics from hyperparameter tuning are in HyperParameterTuning.ipynb though these are trained on less epochs
• The function that calls upon the model created from the other two jupyter notebooks is in MainApp.ipynb
• data/ contains the data processed via the .java code in the same directory from the original dataset used to train the models