The baseline repository stores the the baseline models we used. It includes the training and inference part. The baseline we actually employed in our final project is EfficientNet.ipynb, WaveNet Starter LB 0.52.ipynb, and HMS Resnet1dGRU Inference.ipynb
The Model Combination repository stores the code to run our four model combination approaches and we ensembled them in Final Four ApproachesEnsemble.ipynb. The Inference Features Head Starter and Inference notebook utilized pretrain model. To run these two notebooks, you need to go to the reference link on the top of the notebook and downloaded the model weights first.