Skip to content

Search skills.../

Agent Skill Search Engine

Search

Search
Categories
Occupations

About

About
Privacy
Terms

© 2026 Skills Pool. All rights reserved.

Wandb Integration | Skills Pool

Skill File

Wandb Integration

Integrate Weights & Biases (W&B) as the single system of record for research runs, configs, sweeps, code snapshots, artifacts, checkpoints, and paper-ready analysis. Use when setting up or refactoring experiment tracking, sweep execution, result export, checkpoint resume/eval, or LLM evaluation workflows around W&B.

mahi970 starsApr 8, 2026

Occupation
Categories: Lab Tools

Skill Content

When to Use

Setting up W&B in a new ML or LLM research project
Replacing ad hoc logging with one consistent experiment system
Standardizing run names, config hashes, seed-grouping, and sweep tracking
Adding artifact-based checkpoint storage, reload, resume, or re-evaluation flows
Building scripts that pull runs from W&B and generate paper-quality matplotlib figures

Inputs

Project root and entrypoints for train, eval, sweep, and resume
Main config structure and which fields define the readable run identity
Seed field name and any fields that should be ignored for grouping
W&B project/entity/job_type conventions
Checkpoint policy: latest, best, periodic, final
Whether the project has LLM-specific evals that should use W&B Weave

Core Rule

Treat W&B as the experiment source of truth.

Send configs to wandb.config
Send metrics and logs to

Related Skills

Quick Install

Wandb Integration

npx skillvault add mahi97/mahi97-divan-skills-custom-wandb-integration-skill-md

Download Skill Open repository

Author: mahi97
stars: 0
Updated: Apr 8, 2026
Occupation

On this page

01When to Use 02

wandb.log

Send sweeps through W&B sweeps

Send code provenance and git metadata to W&B

Send checkpoints and other reusable outputs as W&B artifacts

Do not build a parallel logging system unless the project already requires local logs as a cache or failure fallback.

If working from inside divan/, move to the parent project before editing runtime code. Keep the skill in divan/, but put training code, scripts, configs, and tests in the parent project.

Recommended Tracking Contract

Add a small tracking block or equivalent computed fields to every run config.

Inputs

04Recommended Tracking Contract

Data Scientists

Automation Audit Ops

Evidence-first automation inventory and overlap audit workflow for ECC. Use when the user wants to know which jobs, hooks, connectors, MCP servers, or wrappers are live, broken, redundant, or missing before fixing anything.

Github Qa Labels

Label GitHub issues and PRs found during QA testing. Use when organizing QA findings with proper labels.

storybookjs89.7k

Jupyter Notebook

Gamitin kapag humihiling ang gumagamit na gumawa, mag-scaffold, o mag-edit ng mga Jupyter notebook (`.ipynb`) para sa mga eksperimento, eksplorasyon, o tutorial; mas piliin ang mga naka-bundle na template at patakbuhin ang helper script na `new_notebook.py` upang makabuo ng malinis na panimulang notebook.

Tidb Integrationtest Recorder

Use when recording TiDB integration tests under tests/integrationtest and verifying regenerated result files stay minimal and correct.

Quality Nonconformance

Codified expertise for quality control, non-conformance investigation, root cause analysis, corrective action, and supplier quality management in regulated manufacturing.

Hugging Face Trackio

Track ML experiments with Trackio using Python logging, alerts, and CLI metric retrieval.

Data Scientists