for science?if someone would be ablle to provide examples with their answer would be nice.
Clarifying: I did not say that most open source models cannot do porn (most can), just that they cannot do it WELL. A few recent models, like Pony variants, are a little better in terms of poses and the like, but still far away from realistic porn.The only truly uncensored models are those you run yourself. All others will have (either on training or on post-filtering) some restrictions to prevent the generation of some sorts of questionable content (not just porn, and not all porn). That said, most open source models without LORA or other addons will also restrict the kinds of content that can be created due to training.
To be fair that model is over a year old now, there could be better ones out there that can do better hands I haven't looked in a while myself.Interesting that fingers are still an issue for AI. That realistic model looks pretty hot except for the fingers...
Also there are hand-specific LoRAs out in the wild now, so adding them can often fix the issues others have since most models aren't trained to ensure their hands are the best quality.To be fair that model is over a year old now, there could be better ones out there that can do better hands I haven't looked in a while myself.
Yes and no... you can TRY to make any picture, the success you will have will depend on a few factors: the model you are using, the quality of your description of the scene IN TERMS THAT CAN BE APPLIED TO THE MODEL, the quality of the image generation process (which usually means how many tries are you willing to create of your prompt and how many CPU/GPU time are you willing to wait for). So, it is limited, and people who can create good pictures even without AI (e.g. visual artists) usually have a leg up on the generation process, if they are willing to invest time to master it.How does it work? Like can u just make any pictures or?
Sounds like a lot of workYes and no... you can TRY to make any picture, the success you will have will depend on a few factors: the model you are using, the quality of your description of the scene IN TERMS THAT CAN BE APPLIED TO THE MODEL, the quality of the image generation process (which usually means how many tries are you willing to create of your prompt and how many CPU/GPU time are you willing to wait for). So, it is limited, and people who can create good pictures even without AI (e.g. visual artists) usually have a leg up on the generation process, if they are willing to invest time to master it.
It can be, as can any artistic endeavor. However, it is usually still much less work than to create the same result without generative tools, if you are willing to accept the media limitations.Sounds like a lot of work