hfl

/

llama-3-chinese-8b-instruct-v3

like 63

Follow

Joint Laboratory of HIT and iFLYTEK Research (HFL) 425

Hi,I'd like to report a license conflict in hfl/llama-3-chinese-8b-instruct-v3. I noticed that this model was fine-tuned from meta-llama/Meta-Llama-3-8B-Instruct, but it's currently published under the Apache-2.0 license. After taking a look at the META LLAMA 3 COMMUNITY LICENSE AGREEMENT, I found there can be a mismatch between different licensing terms. This inconsistency can make it confusing for people to understand what rules to follow when they use or share the model."

⚠️ Key violations of META LLAMA 3 COMMUNITY LICENSE AGREEMENT:

Clause 1.b.i – Redistribution and Use:
  • ⚠️ No license file included (should contain the META LLAMA 3 COMMUNITY LICENSE AGREEMENT)
  • ⚠️ "Built with Meta Llama 3" is not prominently displayed
  • ⚠️ Model name does not begin with “Llama 3”, which is required for any derivative

Clause 1.b.iii – Required Notice:
  • ⚠️ Missing the following required text in a "NOTICE" file:
    “Meta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.”

Clause 1.iv – Acceptable Use Policy:
  • ⚠️ No mention of Meta’s Acceptable Use Policy, which must be passed on to downstream users

Clause 2 – Additional Commercial Terms:
  • ⚠️ No clarification about the 700M MAU (monthly active users) threshold — making commercial usage ambiguous

On the flip side, Apache-2.0 lets you:

• Use it commercially without asking for extra permission

• Sublicense and redistribute it under more flexible terms

• You don’t have to pass along any non-permissive terms or use restrictions from upstream

This creates a bit of a conflict because the LLaMA 3 license specifically says you can’t sublicense it under more flexible terms and requires downstream users to follow certain use restrictions, which Apache-2.0 doesn’t enforce.

So I'm thinking there might be a licensing conflict here that needs to be sorted out.

🔹 Suggestion:

1. To make sure everything aligns with the LLaMA 3 terms, you might want to tweak the licensing setup a bit, like:

• Maybe include a copy of the LLaMA 3 Community License in the repo or model card

• Include this notice in a “NOTICE” file or the docs:

 > “Meta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.”

• A “Built with Meta Llama 3” note somewhere in the model card could be helpful too

• Maybe a quick note about usage restrictions, especially for folks using it in commercial settings

• A statement clarifying that use of the model must comply with Meta’s Acceptable Use Policy

**2.**Or, we could just drop the Apache-2.0 tag and go with the LLaMA 3 Community License. This could clear up any confusion about redistribution rights and how people can use it downstream.

Hope this helps! 😊 Let me know if you have any questions or need more info.

Thanks for your attention!