Deploying LLM on prem: let's review licenses and highlight open source technologies
summary
A hands‑on review of 40 popular Hugging Face LLM projects finds >50% permissive (MIT/Apache) licenses but notable GPL/AGPL and proprietary API traps that can force source disclosure or royalties. Recommends three concrete practices: avoid GPL/AGPL in deployable images, build on customer premises from private repos, and use SBOM and SPDX to automate license checks. Advises shipping build instructions not monolithic images, generating reproducible builds, and treating license selection as a product decision before development.
AI-written summary · about
chapters (5)
- 0:00Introduction to on-prem LLM licensing
- 3:52Open source license families explained
- 7:13Review of popular AI technologies' licenses
- 10:16Best practices for legal compliance
- 14:36Automating license compliance checks
related talks
01h 38mOpen Source Licensing: Types, Strategies and Compliance
13mBuild license management into your pipelines
14mOpen Source Licensing in a Business World
48mWhat You Need to Know About Open Source - Trust Me, I'm a Lawyer
57mWhat You Need to Know About Open Source—Trust Me, I'm a Lawyer
45mOpen Source distributions in a cloud-native world: from a technical to a legal point of view
30mOpen Source Licensing: Use Versus Abuse
42mOpen Source Licenses: Between Parish Wars and Radicalism