Clarification on license reference to removing content filters?

[ProGamerGov]

I'm wonder if this section of the license is supposed to be included? It appears to say that any removal of the content filters is not allowed under any circumstances. If that is the case, then it's only going to trigger conflict with the community immediately after the release of the weights.

2. All persons obtaining a copy or substantial portion of the Software,
a modified version of the Software (or substantial portion thereof), or
a derivative work based upon this Software (or substantial portion thereof)
must not delete, remove, disable, diminish, or circumvent any inference filters or
inference filter mechanisms in the Software, or any portion of the Software that
implements any such filters or filter mechanisms.

https://github.com/deep-floyd/IF/blob/af64403da0ae2667e5d40670f4014de04bd5c523/LICENSE

[shonenkov]

thank you for this issue! @ProGamerGov

could you give feedback here after release of weights, please

how many false positive and how much complicated making of arts with these kind of filters? this feedback is very welcomed

What about license - we can change filter hyperparameters and find balance for our community - this release under non-commercial license to get feedback and create commercial and free StableIF

[AndreiSva]

This license agreement is not acceptable for supposedly "free" software. It completely undermines the user and depending on what restrictions are put in place, has the potential to unnecessarily restrict creativity and cripple the capabilities of what could be a revolutionary tool.

It is NOT ok for upstream to impose restrictions on what modifications can be made on the software and still call it free. SD has been doing just fine without such a clause. If you're afraid of legal trouble, simply add a disclaimer.

That said I really appreciate what you are doing here having generative AI accessible to more people is good in any situation, even if there are restrictions on what can be done with the code. I just think that there has to be clearer communication on what people can expect from IF, if it is truly free or only open source.

[BadisG]

Welp, guess I'll stick to Stable diffusion then, at least this one doesn't treat me like a child who doesn't know what's "good" or "bad".

[Meatfucker]

Model seems interesting, but the usage restrictions are onerous and will ensure most people do not bother to use it.

[ProGamerGov]

how many false positive and how much complicated making of arts with these kind of filters? this feedback is very welcomed

Content filters are fundamentally art incompatible with free artistic expression, but are useful for when letting people trial systems in a public place.

What about license - we can change filter hyperparameters and find balance for our community - this release under non-commercial license to get feedback and create commercial and free StableIF

@shonenkov So is this license is only meant for the testing phase of the model as part of a way to get users to help you test the filters? If so, then this should be explicitly stated to avoid confusion. It might also be easier to test the filters by holding some sort of contest to find issues then, rather than trying to enforce mandatory content filters.

[AndreiSva]

What about license - we can change filter hyperparameters and find balance for our community - this release under non-commercial license to get feedback and create commercial and free StableIF

@shonenkov It's not about the filters themselves, it's about control. People should be able to individually modify the filters to suit their needs, not be locked into a single immutable filter.

[mcmonkey4eva]

There's a bit of a misunderstanding, allow me to clarify:
this release under non-commercial license to get feedback and create commercial and free StableIF
This means: the initial release of the IF model is under a restricted research-purposes-only license temporarily to gather feedback, and after that feedback is gathered we will be releasing a completely free commercially compatible version.

(I'm from StabilityAI, popping in to help. Apologies for our lovely DeepFloyd team's English, they're not native speakers!)

[comfyanonymous]

That part of the source code license makes it incompatible with free software licenses meaning anyone who wants to implement this in their open source project can't use any of the code in this repo.

It doesn't even make any practical sense why it's included when the diffusers implementation that's linked in this repo has no such restrictions.

[AndreiSva]

@mcmonkey4eva do you mean to say that modifications to the filter will be allowed in the license of the commercially compatible version?

[aleqniko]

what does it mean should not delete delete what else for filters of inferences?
WTF??

[BadisG]

Loooool

https://www.reddit.com/r/StableDiffusion/comments/132fnhq/after_backlash_on_reddit_and_github_stability/

[ProGamerGov]

@AndreiSva @aleqniko

They are still working on the model and seem to want the current test version to not become the dominant version used by the community. Once this first iteration is done testing, then it will be released with an open source license allowing it to spread far and wide with any modifications you desire.

They just made a mistake on communicating this by not including a disclaimer initially, and that led to people fearing the worst: #36

[JorgeCepeda]

I recommend 0BSD, since it's public domain.