GO HERE FOR GENERATE
x1 tensor.art
x1 this model merge dXxjmX-illus-xL-A & dXjm-illustrouX
if you use ComfyUI, you can copy-paste the workflow from the showcase image. replace the encoder with the normal text encoder.
Description
SPECS
UNET BF16
CLIP FP32
VAE BF16
FAQ
Comments (16)
On the 22nd you released x7, then on the 23rd you release x8 but it's in early access? Your models are great, but I wish you explained your thinking a little more.
This model was created and released as a one-shot. As soon as the build is complete and worth trying, I immediately upload it. The Early Access status in version 8 is used to provide a break, so that users continue to use versions 6 and 7 first, while version 8 is still in the early access stage. That’s more or less the idea—I’m a bit bad at explaining things in English. but you definitely get what I mean. 🙂↕️
@DjM_ Ah ok, that's a weird way to do it, but x7 is great anyway so I don't mind.
why 8gb, how 8gb?
You are broke the system, boss)
nah its bf16 + fp32 for clips u can read specs (about this version)
@DjM_ Yes, Its clear, just see it for the first time. thank you a lot😊
@mifink94 yep yep 🫡
@DjM_ I also wanted to ask, is your djMILXLVAE not integrated?
@DjM_ and could you tell me how it works? bf16 + fp32 I means. Or may be site with infoO_O
@mifink94 just try it, i forgot
@DjM_ Yes, I read it, but I still dont understand that it will work better on 4080?(
Sorry for the stupid questions
@mifink94 of course very smoothly