Comments (19)
In clarification and, for the avoidance of any doubt
The license is the license. Apache 2.0 does not contain any ethical restrictions, and that is a good thing.
The base models are licensed under CC BY-SA 4.0 which unambiguously allows use "for any purpose" with "no additional restrictions" beyond requiring that changes to the model have the same license and give credit to StabilityAI:
"Adapt — remix, transform, and build upon the material for any purpose, even commercially.
No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
Ethical "clarifications" would be a step backwards for openness.
from stablelm.
Community Guideline and similar ethical concern documents are actively harmful in open source.
There is, thankfully, no authority to "deal with" ethical concerns and no guidance needed.
The license is the license.
from stablelm.
@MarkSchmidty, I agree.
I think that, as in the free and open source software tradition, the only correct answer to "What ethical restrictions should there be?" is "None".
Everyone can make any ethical statement one wishes on one's own personal blog. The "LICENSE.txt" file is the wrong place to do so.
from stablelm.
Where is the phrase "ethical or appropriate uses" even used in this repository?
It isn't and it should remain that way.
Ethical requirements beyond "don't use it for evil"
Evil is highly ambiguous. I would go as far to say even that requirement is inherently problematic.
Anyway, the licenses clearly allow all uses and are irrevocable. So any kind of ethical statement would be unenforceable and meaningless at best, if not actively harmful.
from stablelm.
"Adapt — remix, transform, and build upon the material for any purpose, even commercially.
No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.Thanks for the details. I thought this model cannot be used commercially based on the license in the HF model hub.
These models are intended to be used by the open-source community chat-like applications in adherence with the [CC BY-NC-SA-4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/) license.
That is the license for the "Tuned" models only, likely because they were trained on datasets of GPT-3/4 output and Stability doesn't want beef with OpenAI. The "Base" model license allows commercial use (with sharealike and attribution).
from stablelm.
Thank you. I've already closed this issue. I'm delighted that 'free' here means 'free' :) .
from stablelm.
You don't need my permission to repeat my words. But I appreciate the sentiment.
from stablelm.
Ethical requirements beyond "don't use it for evil" can be very subjective and ambiguous, which then leads to a potential for actually causing harm. Like for example NSFW bans tend to disproportionately affect the LGBTQ community.
from stablelm.
Where is the phrase "ethical or appropriate uses" even used in this repository?
from stablelm.
I'm actually curious as to why Stability chose to use a unrestricted CC license for StableLM but chose to use a restricted CreativeML Open RAIL license for Stable Diffusion.
from stablelm.
@eiery Maybe they learned from Stable Diffusion release, that it's better to use an unrestricted license?
from stablelm.
Ethical "clarifications" would be a step backwards for openness.
I actually support that viewpoint, and I am encouraged by the strength of the responses here.!
However, I still feel that there should be some kind of guidance document addressing how certain ethical concerns could be dealt with, and how to handle them, even if it's one written with a very open minded and 'liberal' viewpoint, assuming that the overwhelming majority of potential re-users are going to be competent, responsible, and aware of both the potential and pitfalls of the technologies they are creating.
from stablelm.
Ethical requirements beyond "don't use it for evil" can be very subjective and ambiguous, which then leads to a potential for actually causing harm. Like for example NSFW bans tend to disproportionately affect the LGBTQ community.
You make an excellent point here. NSFW bans can disproportionately affect LGBTQ communities, which is why it is my view that tools/models which allow NSFW use, should be upfront about saying this, so that potential re-users can reject "non-free" (GNU sense) tools. Being transparent up front about what you allow is better than burying what you don't in Eula, Terms of Service, or a (Non)-"Sharing" policy, on a website that few people actually read.
from stablelm.
Thank you.. Closing this issue as the consensus here seems to be that such a document is un-necessary and actively harmful.
from stablelm.
However, I still feel that there should be some kind of guidance document addressing how certain ethical concerns could be dealt with, and how to handle them,
What ought to be done is a matter of opinion anyone can discuss, regardless of whether they are publishing models or not.
There is really no reasons for statements like that to be part of this repo. Anyone, including of course anyone working at StabilityAI, can make them outside of technical places, such as on blogs and the like.
from stablelm.
"Adapt — remix, transform, and build upon the material for any purpose, even commercially.
No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.
Thanks for the details. I thought this model cannot be used commercially based on the license in the HF model hub.
These models are intended to be used by the open-source community chat-like applications in adherence with the [CC BY-NC-SA-4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/) license.
from stablelm.
Models under the CC BY-NC-SA license cannot be used commercially. Models under CC BY-SA can, but of course there is a ShareAlike clause.
from stablelm.
Community Guideline and similar ethical concern documents are actively harmful in open source.
Do you mind if I use that quote elsewhere (do you have an appropriate attirbution I could use)?
Restrictions on use imposed by other AI models/datasets are EXACTLY why there need to be "Free" (GNU sense) models :)
from stablelm.
You don't need my permission to repeat my words. But I appreciate the sentiment.
Thanks, if you want to comment in the linked issues directly, feel free :) . Encouraging 'free' (Gnu sense) is a good thing.
from stablelm.
Related Issues (20)
- Any advice how to train a model of a different language? HOT 2
- What's the proper way to implement chatting feature? HOT 6
- RLHF training code for StableVicuna open sourced? HOT 1
- StableVicuna does not stop dialog speaking, probably until max_new_tokens. HOT 3
- loss not decreasing with deepspeed HOT 1
- Training Script stablity 3B and 7B HOT 6
- Unclear tokenizer class HOT 2
- Cannot run demo HOT 2
- fairyfloss HOT 2
- process killed HOT 4
- License unclear HOT 8
- Is it normal to take a long time ( about 15min )to generate an answer? HOT 1
- How to expand the sequence length of llama? HOT 1
- Consider using OpenAI Evals
- The output is the same as the input. HOT 1
- Is this project abandoned? HOT 4
- Stability AI
- Hello, how to convert the statityai/tablelm-base-alpha-3b to ggml format HOT 1
- Target modules ['query_key_value', 'dense', 'dense_h_to_4h', 'dense_4h_to_h'] not found in the base model. Please check the target modules and try again. HOT 2
- OSError: stabilityai/stablelm-base-alpha-3b-v2 does not appear to have a file named pytorch_model.bin, tf_model.h5, model.ckpt or flax_model.msgpack. HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from stablelm.