An SDXL update of EnvyPoodaMix. This checkpoint straddles the line between anime and western comics.
Shout out to @DuskFallCrew for the original 1.5 Pooda-Beep Mix checkpoint!
Description
FAQ
Comments (10)
Umm, you only posted the sdxl vae file. 319 mb seemed a bit low for an sdxl checkpoint. 😋
Yeah, my bad. It's fixed now. :)
Damnit now i gotta make Poodabeep XL
For some reason, the model doesn't work. When creating a picture it gives the following error: Input type (float) and bias type (struct c10::Half) should be the same
Am I doing something wrong?
Hey Envy, does it look like these XL models may be able to run on lower spec PCs?
SD basic with all models on this site are able to run on 4GB VRAM, like my pc, without much issues.
But XL can't. :(
There's talk that they "CAN" in theory but you're talking 1-5 minute generation times, and Cuda errors likely. Envy has some of their models on Tensor, and that's an easy place to generate SDXL content :)
I can run XL with 4GB, but without refiner by using ComfyUI
@duskfallcrew webservices like the one you mentioned have very little in the way of customization however. I similarly to Envy, although not to the same degree, like to spend time refining prompts, loras, control nets, etc.
@Shio_N ty bro, gonna check it out
@colcaboom Tensor has control net, Adetailer and hires fix - and tons of loras, HOWEVER i do understand that generation services like that are for more beginners rather than pro. I'm only commenting about Tensor because it's an option, not because i actually care enogh to shill my link (They got me to bring a lot of my stuff over and now i ahve too many ponts and can't conver them to money XD)


















