Which tracking tools does this skill support?

It provides implementation patterns and best practices for MLflow, Weights & Biases (wandb), Neptune, Comet, and DVC.

Can I use this for hyperparameter tuning?

Yes, it includes specific patterns for managing hyperparameter sweeps, including configuration for Bayesian, grid, and random search methods.

How does this skill help with experiment reproducibility?

It offers detailed checklists and code patterns for pinning git commits, locking environment dependencies, and versioning datasets to ensure training results can be replicated exactly.

What is the recommended project structure for ML experiments?

The skill suggests a clear hierarchy separating baseline experiments, feature engineering, and hyperparameter tuning, alongside versioned data and model staging folders.

ML Experiment Tracker

Name: ML Experiment Tracker
Author: organvm-iv-taxis

byorganvm-iv-taxis

•

Ciencia de Datos y ML

Streamlines machine learning experiment logging, model versioning, and reproducibility using industry-standard tools like MLflow, Weights & Biases, and DVC.

This skill empowers developers and data scientists to maintain a rigorous, systematic approach to machine learning development by implementing standardized tracking and versioning protocols. It provides comprehensive guidance on logging hyperparameters, metrics, and artifacts across popular platforms like MLflow and Weights & Biases, while integrating DVC for robust data and pipeline versioning. By following these established patterns, users can ensure full experiment reproducibility, simplify model comparisons, and maintain a clear lineage from raw data to production-ready models.

Características Principales

01Model registry workflows for staging and production deployments

02Comprehensive reproducibility checklists for code, data, and environments

03Standardized logging for hyperparameters, metrics, and training artifacts

04Data versioning and pipeline management using DVC

05Integration patterns for MLflow, Weights & Biases, and Neptune

062 GitHub stars

Casos de Uso

01Implementing data version control to track the exact dataset lineage for specific models

02Managing hyperparameter sweeps and comparing model performance across multiple training runs

03Establishing a formal model registry to manage the lifecycle of candidate models from dev to prod

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add organvm-iv-taxis/a-i--skills ml-experiment-tracker

For use in Claude.ai and ChatGPT

Características Principales

01Model registry workflows for staging and production deployments

02Comprehensive reproducibility checklists for code, data, and environments

03Standardized logging for hyperparameters, metrics, and training artifacts

04Data versioning and pipeline management using DVC

05Integration patterns for MLflow, Weights & Biases, and Neptune

062 GitHub stars

Casos de Uso

01Implementing data version control to track the exact dataset lineage for specific models

02Managing hyperparameter sweeps and comparing model performance across multiple training runs

03Establishing a formal model registry to manage the lifecycle of candidate models from dev to prod

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add organvm-iv-taxis/a-i--skills ml-experiment-tracker

For use in Claude.ai and ChatGPT