Model License Labyrinth: Llama vs Gemma vs Qwen vs Mistral — 'Derivative Work' Debate
Which license when publishing FT model? How does base model's license affect **derivative work**? Llama Community License v3 (>700M MAU restriction), Gemma ToS (responsible use), Qwen2 Apache 2.0 (most flexible), Mistral Research vs Apache (model-specific), OpenAI ToS (output restriction). Cookbook decision matrix.
Şükrü Yusuf KAYA
24 min read
Intermediate
1. Lisans Matrix#
| Base Model | Lisans | Derivative kullanım | Kısıtlar |
|---|---|---|---|
| Llama 3.x | Llama Community v3 | OK, dağıtılabilir | >700M MAU kullananlar Meta'dan izin alır |
| Gemma 3 | Gemma ToS | OK | "Prohibited Use Policy" |
| Qwen 2.5/3 | Apache 2.0 | OK kısıtsız | Lisans bildirimi |
| Mistral 7B v0.3, Mistral Small 3, Codestral Mamba | Apache 2.0 | OK kısıtsız | — |
| Mistral Large 2, Codestral 22B | Mistral Research | hayır (commercial) | research only |
| DeepSeek-V3, R1 | MIT | OK kısıtsız | — |
| Phi-4 | MIT | OK kısıtsız | — |
| Aya Expanse | CC-BY-NC | hayır (commercial) | research only |
| Cohere Command-R | CC-BY-NC | hayır (commercial) | research only |
Cookbook'un TR commercial production kuralı: Apache 2.0 veya MIT modeller. En esnek: Qwen 2.5/3, DeepSeek-V3, Phi-4.
✅ Teslim
- Kendi base modelinin lisansını yeniden oku. 2) Derivative work formatı belirle. 3) Sonraki ders: 18.4 — Veri Lisans Zinciri.
Yorumlar & Soru-Cevap
(0)Yorum yazmak için giriş yap.
Yorumlar yükleniyor...
Related Content
Part 0 — Engineering Foundations
Welcome to the Fine-Tuning Cookbook: System, Stage Taxonomy, and the Reproducibility Contract
Start LearningPart 0 — Engineering Foundations
Reproducibility Stack: Seeds, cuDNN Flags, and Deterministic CUDA — End the 'Works on My Machine' Problem
Start LearningPart 0 — Engineering Foundations