skills.rest
qcmuu
qcmuuCOMMUNITY
Explore all skills by this author →
GitHub

weights-and-biases

Track experiment runs, log metrics, and manage hyperparameters with Weights & Biases.

Updated May 2, 2026
Search Tags:#mlops#experiment tracking#wandb#model registry#artifact versioning#hyperparameter sweeps#training visualization
One-click install
npx skills add https://github.com/qcmuu/AI-Research-Skills --skill weights-and-biases-qcmuu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: weights-and-biases
Source: https://github.com/qcmuu/AI-Research-Skills/tree/main/13-mlops/weights-and-biases
Command: npx skills add https://github.com/qcmuu/AI-Research-Skills --skill weights-and-biases-qcmuu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires wandb, and includes references (resource) components.

What problem does it solve?

Weights & Biases (W&B) helps you stop losing context across training runs by centralizing experiment tracking, metrics visualization, hyperparameter sweeps, and model/artifact management.

Core Features & Use Cases

  • Experiment Tracking & Real-Time Dashboards: Log scalar metrics, media, tables, and system stats to monitor training progress as it happens.
  • Configuration & Metric Management: Automatically capture hyperparameters and correlate them with outcomes across runs.
  • Hyperparameter Sweeps: Run automated searches (grid/random/bayesian) to find better hyperparameters efficiently.
  • Artifacts & Model Registry: Version datasets, checkpoints, and models with lineage, aliases (latest/best/production), and governance-friendly promotion flows.

Use Case: You iterate on a PyTorch or Hugging Face training script and need to compare many runs, identify the best configuration by validation metrics, and register the resulting model for downstream evaluation or deployment.

Quick Start

Tell an AI assistant to set up a W&B run for your training script by logging your learning rate and train/validation losses each epoch, saving checkpoints as W&B artifacts, and (optionally) defining a sweep that maximizes val/accuracy.

Frequently Asked Questions about weights-and-biases

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I track experiment runs and log metrics during model training?▼

To track experiment runs, initialize a W&B run and use wandb.log to report scalar metrics, media, and system stats. This automatically captures hyperparameters and generates real-time dashboards to monitor training progress.

What is the best way to run automated hyperparameter sweeps?▼

Automated hyperparameter sweeps use grid, random, or Bayesian search to efficiently find better configurations. You define the sweep parameters and optimization metric, and the system orchestrates runs to maximize validation accuracy.

How does artifact versioning work for datasets and model checkpoints?▼

Artifact versioning logs datasets and checkpoints with lineage and aliases like latest, best, or production. You create artifacts during training and download specific versions for downstream evaluation or promotion flows.

Can I use Weights & Biases with PyTorch or Hugging Face training scripts?▼

Yes, you can integrate W&B with PyTorch or Hugging Face scripts to log learning rates and validation losses each epoch. This allows you to compare many runs, identify the best configuration, and register the resulting model.

How do I compare multiple training runs to identify the best configuration?▼

You compare multiple training runs using organized W&B dashboards that correlate captured hyperparameters with validation outcomes. This centralized experiment tracking visualizes results across runs to identify the best performing configuration.

Related Skills

Explore More Software Engineering Skills→
terrylicaterrylica
62

mlflow-python

Log MLflow metrics, parameters, and artifacts using the Python API.

Community
Advanced
kevinnftkevinnft
13

weights-and-biases

Track ML experiments, sweeps, and artifacts using Weights & Biases.

Community
Advanced
Z43LZ43L

weights-and-biases

Log ML experiments, hyperparameter sweeps, and model artifacts with wandb.

Community
Intermediate
cxnaivecxnaive

weights-and-biases

Tracks machine learning experiments, hyperparameter sweeps, and model artifacts.

Community
Advanced