Home Artists Posts Import Register

Content

https://civitai.com/user/FilthyDegenerate

Trying to squeeze (pun intended) as much realism from Vendo's as possible, I managed to improve it a bit after fixing my fucked up positive and negative prompts as one specific tag was preventing me to increase the quality, also let me remind you, I use this shit daily non-stop (just like I did with blender, remember?), I know my stuff, no other 'realistic' pony model has good prompt adherence like this one, most of these other models sacrifice consistancy in prompts with quality, and that is bad, real bad, why is it bad? the more LoRa's you have to use the more the quality of your image is degraded because LoRa's aren't trained with realistic images in mind, rarely with see realistic LoRa's on civitai.

an example is ponyrealism v11, it dislikes asians and blacks lol, the asian features are more akin to the ones in the early days of 1.5 that the eyes aren't that 'closed'.

Most of the top rated models on civitai are very very questionable lol.

Files

Comments

joe cool

this is pretty incredible realism imo would love to see more if possible

FilthyDegenerate

Is called 2.5 I think, not realistic, as 'realism' or something convincing can only be achieved on older already mature models like Juggernaut or Epic Original Sin, new Pony models aren't near close that.