Weights And Biases

This is the standard toolkit for tracking machine learning experiments and keeping your sanity when you're running dozens of training runs. It automatically logs metrics, hyperparameters, and system stats to a web dashboard where you can compare runs side by side. The sweep functionality is legitimately useful for hyperparameter optimization, supporting grid, random, and Bayesian search strategies. You can also version datasets and models as artifacts with full lineage tracking. The integration code is minimal, usually just a wandb.init() and wandb.log() calls in your training loop. Works seamlessly with PyTorch, TensorFlow, and HuggingFace. If you've ever lost track of which learning rate produced your best model, this solves that problem.

skills.sh installs: 1.1K
GitHub stars: 14.4K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add orchestra-research/ai-research-skills --skill weights-and-biases --agent claude-code

Installs into .claude/skills of the current project.

Files

Weights & Biases: ML Experiment Tracking & MLOps

When to Use This Skill

Use Weights & Biases (W&B) when you need to:

Track ML experiments with automatic metric logging
Visualize training in real-time dashboards
Compare runs across hyperparameters and configurations
Optimize hyperparameters with automated sweeps
Manage model registry with versioning and lineage
Collaborate on ML projects with team workspaces
Track artifacts (datasets, models, code) with lineage

Users: 200,000+ ML practitioners | GitHub Stars: 10.5k+ | Integrations: 100+

Installation

# Install W&B
pip install wandb

# Login (creates API key)
wandb login

# Or set API key programmatically
export WANDB_API_KEY=your_api_key_here

Quick Start

Basic Experiment Tracking

import wandb

# Initialize a run
run = wandb.init(
    project="my-project",
    config={
        "learning_rate": 0.001,
        "epochs": 10,
        "batch_size": 32,
        "architecture": "ResNet50"
    }
)

# Training loop
for epoch in range(run.config.epochs):
    # Your training code
    train_loss = train_epoch()
    val_loss = validate()

SKILL.md on GitHub