wandb-tracker

Log ML experiment metrics, artifacts, and hyperparameters to Weights & Biases.

Updated Feb 17, 2026

Search Tags:#mlops #observability #hyperparameter tuning #experiment tracking #weights and biases #artifact versioning

Ready to run

npx skills add https://github.com/Rachasumanth/text2llm001 --skill wandb-tracker

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: wandb-tracker
Source: https://github.com/Rachasumanth/text2llm001/tree/main/workspace/skills/wandb-tracker
Command: npx skills add https://github.com/Rachasumanth/text2llm001 --skill wandb-tracker

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill streamlines the process of tracking machine learning experiments, ensuring reproducibility, and facilitating easy comparison of training runs and hyperparameter tuning results.

Core Features & Use Cases

Experiment Tracking: Log metrics, hyperparameters, and artifacts for ML training runs.
Artifact Versioning: Version control for model checkpoints, tokenizers, and datasets.
Hyperparameter Tuning: Facilitates side-by-side comparison of runs to identify optimal configurations.
Use Case: When training a new deep learning model, use this skill to automatically log training/validation loss, learning rate, and GPU utilization to Weights & Biases, allowing you to visualize progress and compare different model architectures.

Quick Start

Use the wandb-tracker skill to initialize a new W&B run for the 'image-classification' project with the entity 'my-org'.

Frequently Asked Questions about wandb-tracker

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

FAQPage Schema

How do I track machine learning experiment metrics and hyperparameters with Weights and Biases?▼

To track machine learning experiment metrics and hyperparameters with Weights and Biases, initialize a new run for your project to automatically log training loss, learning rate, and GPU utilization for visual analysis.

What is artifact versioning for model checkpoints and datasets?▼

Artifact versioning for model checkpoints and datasets is the process of applying version control to machine learning artifacts. This ensures full lineage tracking and reproducibility for your training runs by maintaining historical versions of tokenizers and model states.

How do I compare hyperparameter tuning runs side-by-side?▼

To compare hyperparameter tuning runs side-by-side, log your metrics and configurations to Weights and Biases during training. This automatically generates comparative analysis visualizations to help you identify the optimal model configurations.

Do I need a WANDB_API_KEY environment variable to log training metrics?▼

Yes, you need a WANDB_API_KEY environment variable to log training metrics. This key is required for authenticating with the Weights and Biases API and establishing access to your organization's experiment tracking workspace.

Can I use experiment tracking to log GPU utilization during model training?▼

Yes, you can use experiment tracking to log GPU utilization during model training. The system automatically records system-level metrics alongside training and validation loss to provide comprehensive observability for deep learning development.

What are the limitations of using Weights and Biases for experiment observability?▼

A key limitation of using Weights and Biases for experiment observability is the strict dependency on the WANDB_API_KEY environment variable. Without this authentication prerequisite, the skill cannot initialize runs or synchronize logging data to the platform.