This project benchmarks XGBoost and PyTorch models for binary classification across serial CPU, parallel CPU, and GPU setups. Parallel XGBoost achieved the fastest training time without loss in accuracy (~52%), while PyTorch on GPU delivered the best F1 Score (0.4592), highlighting trade-offs between speed and performance on small tabular datasets.
data-science machine-learning deep-learning neural-network pytorch xgboost gpu-acceleration parallel-processing binary-classification serial-processing
-
Updated
May 13, 2025 - Jupyter Notebook