Pawpularity

Kaggle Competition Bronze Medal 🥉 (205th out of 3537 teams)

Machine Learning course project @ Tsinghua University

Authors

Armando Fortes & David Pissarra

Overview

Nowadays, millions of stray animals are alone on the streets or living in shelters, waiting for a new home. With the aim of better understanding the impact of pet photos and their features, PetFinder.my - Malaysia's leading animal welfare platform - designed a metric which measures pet photo popularity - Pawpularity - and hosted a related Kaggle competition.

Follows the architecture of our best performing solution for the competition:

Further details on our solution can be found in the report.

Built With

• Swin Transformer (Images)
• XGBoost (Metadata)
• Optuna (Ensemble Model tuning)