Home▸Data Science▸GradNorm: Loss Weighting for Multi-Task Learning

GradNorm: Loss Weighting for Multi-Task Learning

Watch three task heads train through one shared trunk while Equal, Uncertainty, or GradNorm loss weighting decides how much each task's gradient pulls on the shared weights — with live loss curves, per-task weights, and gradient norms.

Data Science3DAdvanced60 FPS

A shared trunk feeds three task-specific heads, and this simulator lets you switch between the three standard ways a real multi-task network decides how hard each task is allowed to pull on those shared weights: naive equal weighting, Kendall's learned-uncertainty weighting, and Chen et al.'s GradNorm, which actively balances each task's gradient magnitude against its relative training pace. Colored particle streams flow from each task head back into the trunk with a size and speed set by that task's live gradient norm, while side panels track every task's evolving loss weight and gradient norm step by step — so you can watch, in real time, a hard task get starved under equal weighting and then rescued the moment GradNorm or uncertainty weighting takes over.

⚙ Under the hood

Watch three task heads train through one shared trunk while Equal, Uncertainty, or GradNorm loss weighting decides how hard each task's gradient pulls on the shared weights, with live loss, weight, and gradient-norm readouts.

multi-task learninggradnormloss weightingdeep learningneural networksmachine learning

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install