There are a lot of things that don't appear in ELO scores. For one, they will not reflect that you cannot prompt women's faces in Flux. We can only speculate why.
There are a lot of things that don't appear in ELO scores. For one, they will not reflect that you cannot prompt women's faces in Flux. We can only speculate why.
It is actually making it harder to use the technology to represent women characters, which is so ironic. That said, I could just lEaRn tO dRaW or pAy aN aRtIsT right? The discourse around this is so shitty.
In any case, I believe that you're not generating lewd photos (I hope you don't blame me too much for being surprised though, it's _incredibly_ common). I still imagine the reason you're having trouble though is _because_ such descriptions are relatively adjacent enough for whatever censoring measures they used to nerf lewd outputs.
Censorship is kinda bullshit. But, it's a very expensive toy that they have given away for free (the lesser models, anyway). You can use a LORA (or hell, train your own presumably million-dollar model) to remove these restrictions.
It's more frustrating when you're paying for it and it's behind an API you have no control over. Which, I realize BFL is doing for their pro model - but this isn't anywhere near as egregious as, say, the level of censoring done behind the API for the DALL-E models. And, at the very least it is understandable from a _company_ perspective. They have both legal and public relations concerns to manage. To say that they should just ignore those facets of running a company feels a little entitled to me.
I feel like AI should just be treated as fair use as long as its not 100% blatantly a literal clone of the original work.
Ideogram and Flux both have their own broad set of limitations that are non-technical and unpublished. IMO they are not really motivated by legal concerns, other than the lack of transparency itself.
So maybe the issue is that transparency, and that the hazy legal climate means no transparency. You can't go anywhere and see the detailed list of dataset collection and captioning opinions for proprietary models. Open Model Initiative, trying to make a model, did publish their opinions, and they're not getting sued anytime soon. However, their opinions are an endless source of conflict.