I asked a generator to make me a “queen monkey in a purple gown sitting on a throne” and I got maybe two pictures of actual monkeys. I even tried rewording it several times to be a real monkey, described the hair and everything.
Stable diffusion online version, several weeks ago. Might not be the same situation anymore, idk how often that stuff gets updated, and I’m not able to test it at the moment.
It’s also possible that some sort of “sticky idea” got into its head and made it start generating it that way after it did one like that. I’ve heard that sort of thing isn’t uncommon.
To be clear, stable diffusion isn’t one model, it’s the generation platform. From there, you have models that sit on top of it. Online generators can use any model, depending on how they’re set up. Each model includes different training data, meaning different results from the same prompts, sometimes vastly.
It’s a bit like driving somewhere, having someone ask how you found the place, and saying your phone. Technically a correct answer, but they’re probably looking for more specific answers, like GPS, or a map. Not trying to nit-pick, just giving a bit of information.
Significant racist bias is an understatement.
I asked a generator to make me a “queen monkey in a purple gown sitting on a throne” and I got maybe two pictures of actual monkeys. I even tried rewording it several times to be a real monkey, described the hair and everything.
The rest were all women of color.
Very disturbing. Pretty ladies, but very racist.
deleted by creator
Stable diffusion online version, several weeks ago. Might not be the same situation anymore, idk how often that stuff gets updated, and I’m not able to test it at the moment.
It’s also possible that some sort of “sticky idea” got into its head and made it start generating it that way after it did one like that. I’ve heard that sort of thing isn’t uncommon.
To be clear, stable diffusion isn’t one model, it’s the generation platform. From there, you have models that sit on top of it. Online generators can use any model, depending on how they’re set up. Each model includes different training data, meaning different results from the same prompts, sometimes vastly.
It’s a bit like driving somewhere, having someone ask how you found the place, and saying your phone. Technically a correct answer, but they’re probably looking for more specific answers, like GPS, or a map. Not trying to nit-pick, just giving a bit of information.
Stable diffusion is a generative technology, not a model