Search Results
6/18/2025, 11:45:32 PM
>>2975020
>at least used an anime finetuned model if you want to use 1.5
This is an anime finetuned model. The sample images are so much better than my results it isn't even recognizable as the same model.
> in terms of sdxl, your card probably only supports fp32 (32bit floats) if it doesn't load. usually, sdxl uses around 5-6gb of vram with fp16, but if your card doesn't support fp16 it will double the amount of vram needed (minus the clip model most likely since you're likely not fitting into vram)
The chart I'm looking at says that it does support fp16. Not that I know particularly well what that means.
I get this error on the image when I try to run autismmixSDXL.
Also, I remember that I actually did run XL models like Illustrius and autismmix before on SD NEXT, but it doesn't work now for some reason and I can't figure out what's different.
>at least used an anime finetuned model if you want to use 1.5
This is an anime finetuned model. The sample images are so much better than my results it isn't even recognizable as the same model.
> in terms of sdxl, your card probably only supports fp32 (32bit floats) if it doesn't load. usually, sdxl uses around 5-6gb of vram with fp16, but if your card doesn't support fp16 it will double the amount of vram needed (minus the clip model most likely since you're likely not fitting into vram)
The chart I'm looking at says that it does support fp16. Not that I know particularly well what that means.
I get this error on the image when I try to run autismmixSDXL.
Also, I remember that I actually did run XL models like Illustrius and autismmix before on SD NEXT, but it doesn't work now for some reason and I can't figure out what's different.
Page 1