Weights and Biases

Definitions

1. A machine-learning tooling platform commonly used for experiment tracking, model monitoring, and training visualization.

   In plain English: A popular tool for tracking ML experiments and training runs.

   Example: "Every run was logged to Weights and Biases so the sweep stayed comparable."

Related Terms

• Function Calling
• Mixed Precision Training
• MLflow
• Model Architecture
• Model Card Detail
• Model Compression
• Model Distillation
• Model Ensemble