Machine Learning Icons

This collection follows the machine learning workflow from start to finish: data collection, labeling, splitting, feature engineering, training loops, loss curves, gradient descent, hyperparameters, model types (neural networks, decision trees, clustering, regression, transformers), evaluation charts such as confusion matrices and ROC curves, inference, GPUs, pipelines, experiment tracking and drift detection. The 150 icons come in line and duotone styles on a 24 px grid, with the duotone accent controlled by one Figma variable.

Data scientists preparing talks, ML platform teams, educators and technical writers get specific symbols for concepts that generic tech sets skip. Clear icons for training, evaluation and deployment make course materials, documentation and product screens easier to scan.