AI Model Pre-Release Checklist
Check everything that has to be settled before a model goes public.
Releasing an AI Model
Preparing and releasing a trained AI model on Hugging Face or a similar hub
Releasing source code instead?
This page covers releasing a trained AI model (its weights). For source code, see Releasing Open Source. If you are consuming an external open source model rather than publishing one, see AI Model Licenses.
Releasing an AI model follows most of the same steps as releasing source code. You obtain organizational approval, confirm you have the right to publish, remove sensitive information, and assign someone to support the project afterwards. For those shared steps, follow the release process and the release rules.
This page covers only what differs because the artifact is a model.
What differs from a code release
| Aspect | Source code | AI model |
|---|---|---|
| What you publish | Source code | Weight files and a model card |
| Licensing | One license for the code | Model license and training dataset licenses, judged separately |
| Documentation | README, contribution guide | Model card covering intended use, limits, bias, evaluation |
| Sensitive material | Code and commit history | Also personal data and copyrighted works inside the training data |
| Regulation | Export control (ECCN) | Also the documentation duties of the EU AI Act and Korea’s AI Framework Act |
Training datasets are where teams most often get stuck. Even with the model license settled, the license of the data you trained on may restrict redistribution or commercial use. Check the two separately.
Summary
- Obtain internal approval and request an OSRB review (stage A of the release process).
- Confirm the license of both the model and every training dataset.
- Work through the pre-release checklist.
- Write the model card. It is the document your users will read first.
- If you need to prepare for regulation, check your documentation with an AI SBOM.
- Publish the repository and operate it (stage D of the release process).
Decision flow
| Question | Next step |
|---|---|
| Do you have the right to redistribute the weights (check the base model license)? | If unclear, request a review first |
| Do the training dataset licenses permit redistribution? | If restricted, decide whether to publish the model without the dataset |
| Does the training data contain personal data or third-party copyrighted works? | If so, consult Legal and the privacy team before publishing |
| Is someone assigned to answer questions after release? | If not, plan for that first |
Getting a review before you publish
You can have the model checked while the repository is still private. Push the model privately and share read access with the OSRB, and we will report what is missing and how to fill it. Grant access with a token scoped to that single repository rather than one that opens your whole account.
If your organization enforces a token approval policy, the token stays pending until an administrator approves it. When you have a fixed release date, start the token request early.
Related pages
- Release process — the shared path from approval to operation
- Release rules — what to satisfy before publishing
- Sensitive information checklist — cleaning code and commit history
- AI Model Licenses — RAIL, Llama and similar licenses
Contact
For questions and review requests about releasing an AI model, contact the OSRB (opensource@sktelecom.com).
Writing the Model Card
What a model card is and which fields you need to fill in.
AI SBOM and Regulation
Build an inventory of your model and check how far its documentation goes.
Feedback
Was this page helpful?
Glad to hear it! Please tell us how we can improve.
Sorry to hear that. Please tell us how we can improve.