CS5242 Kaggle Project

Liu Mingzhe - [email protected]
Jeremy Tan - [email protected]

Installation and Usage

1. Install dependencies. pip install -r requirements.txt

2. Save data file in /data folder and extract there.
   The expected filepath is:

   data/
   
     nus-cs5242/
   
   	test_image/
   	train_image/
   	sample_submission.csv
   	train_label.csv
   

3. Run train script.
   This will augment the data, train the model, and output predictions in /predictions folder.
   python train.py
   On the terminal, the expected output should look like this:

Of course, running the full 100 epochs will give a better result.

Expected Output of predictions

The predictions will be saved in the predictions\<timestamp> folder in csv.
The model will likewise be saved in the models\<timestamp> folder.

Repo Structure