GREAT SCOTT! Vivid insanity is here!
===========================================================================
KR34M - 11/25 - Let's give FLUX/KREA the DR34M makeover. This checkpoint excels at photorealism and abstract artwork. We hope you enjoy what's likely our last FLUX release!
Note: NF4 is actually K4_S in GGUF format. FP8 is actually Q8_0, also GGUF!
We recommend the BF16 variant + T5 BF16 if you have the VRAM.
Note: Use LoRA with C4PACITOR at slightly lower than normal weights for best results and to avoid anatomy blending.
===========================================================================
C4PACITOR models are created by DR34MSC4PE with all the trappings you've come to expect as well as some cool bonuses:
Enhanced realism and photo-realistic concepts, trained on high quality datasets with the latest techniques.
Specific anime/illustration tuning to introduce some lost artistic concepts
NSFW tuned and capable of realism and artistic/anime images. Female anatomy is well represented with additional fine tuning in the works.
Exceptional performance with character/other/stacking Lora
Like our work? Buy us a coffee: https://ko-fi.com/dr34msc4pe
===========================================================================
DR34MSC4PE is
c0ur4ge (training/qa/inference code) /
eraser851 (training/captions/tooling code/data/qa)
===========================================================================
Recommendations by base model:
dev - (d_v1/d_v2/a_v)
===========================================================================
CFG: 3-7
Steps: 22-60
Sampler: RES_2M
Scheduler: Beta/Beta57/Bong Tangent
========================================
Hyper8/L16HT - (x-series)
===========================================================================
CFG: 3-6
Steps: 16-32
Sampler: DEIS
========================================
schnell - (s_v0/s_v1):
===========================================================================
Note: 100% Schnell Base
Steps: 2-14
Sampler: DEIS
===========================================================================
Description
Meet, x_v1 - I'm sure you're going to be fast friends. :O) Note - We're aware of the tendency for excessive blushing, this will be balanced out in future versions but is, often, pretty nice...
FAQ
Comments (13)
Hello. I couldn't run it in my current flux workflow. Do you have a workflow you can recommend? Thanks. (Error: Prompt execution failed. TypeError: Failed to fetch)
It should work in any default. What node are you using to load the checkpoint?
@c0ur4ge I tried it on the default workflow used for Flux.dev and it worked. I think it was related to node, thank you.
bab
ill try this
Surprised and happy to see NF4 here, as I've never moved from Flux Dev NF4 so far due to GPU. To confirm though - since this is only 6.3 GB, its not an AIO NF4, like the Dev one, correct? It needs a clip and vae?
That’s right!
awesomely
Is there a plan to launch it to Flux 1 Kontext?
Krea and Kontext have test variants I need to work on. WAN has taken over my time lately but I’m taking a break there for a bit so I’d like to get these out soon!
First flux model that does what i actually want it to for the most part. Could be trained on more artistic concepts like fantasy characters but otherwise it cooperates with loras very nicely.
with that said, when i say most part, i think it would be much better as a q8 gguf instead of an fp8.
It can and does. I will LIKELY replace the FP8 with the GGUF this evening tbh. Folks that wanna quant the BF16 to FP8 can do that or enjoy a limited edition cut of the model.
Look for a range of quants on a different card (or at least Q8 here instead of FP8) shortly!
Update: Q8 is up and has replaced FP8.
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.