# Design exercise: Open-Weight Models

> Model cards, licenses, weights, tokenizers, chat templates, adapters, provenance, and reproducible packaging. Implement designOutline() returning non-empty values for: license, provenance, runtimeFit. Each value must name a concrete mechanism or decision.

Canonical: https://learn.significanthobbies.com/curriculum/drills/practice-open-weight-models

Model cards, licenses, weights, tokenizers, chat templates, adapters, provenance, and reproducible packaging. Implement designOutline() returning non-empty values for: license, provenance, runtimeFit. Each value must name a concrete mechanism or decision.

## The mechanism

Open weights provide inspectable parameters, not automatically open data or unrestricted rights. Selection must include license, provenance, tokenizer, template, eval, and hardware fit.

## Expected evidence

A design outline with license, provenance, runtimeFit plus an explicit failure mode or trade-off.

## Hints

- Name the failure mode before selecting the mechanism.

## Sources and practice

- [The Llama 3 Herd of Models](https://arxiv.org/abs/2407.21783)
- [DeepSeek-V3 Technical Report](https://arxiv.org/abs/2412.19437)
- [Rethinking open source generative AI: open-washing and the EU AI Act (FAccT '24)](https://facctconference.org/static/papers24/facct24-120.pdf)
- [OLMo: Accelerating the Science of Language Models](https://arxiv.org/abs/2402.00838)
- [Hugging Face Model Cards](https://huggingface.co/docs/hub/model-cards)
- [Study Open-Weight Models](https://learn.significanthobbies.com/curriculum/concepts/open-weight-models)
- [Try the interactive drill](https://learn.significanthobbies.com/drills/practice-open-weight-models)
- [All practice briefs](https://learn.significanthobbies.com/curriculum/drills/)