CNN Depth and Width Study
Controlled CIFAR-10/100 depth-vs-width study under a fixed parameter budget.
Applied ML research, data tools, apps, and chess projects with live demos and papers.
Controlled CIFAR-10/100 depth-vs-width study under a fixed parameter budget.
Benchmarked RF, MLP, and CNN classifiers on PathMNIST histology tiles.
Tested denoising and augmentation before histology classification.
Live recession-risk dashboard with yield-curve signals and validation views.
Graph toolkit for communities, centrality, influence, and link prediction.
Optimizer visualizer for learning-rate behavior and loss landscapes.
Bushfire-risk map demo pairing model outputs with response geography.