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
C4P + KREA = KR34M - Need I say more?
FAQ
Comments (11)
KR34M_v1 is amazing! I'm getting good results with a basic 20 step / euler / simple workflow, but what are your recommended settings for it?
haven't attempted to use flux in a long while might start with this.
Please let me know what you think! :)
Can 4090 load a full 22GB model?Cuz the gguf is too slow(2 output= 65s) and can`t accelerate by sage attention KJ
Should be able to - recommend offloading the TENC and VAEs and/or loading it with fp8 as the dtype. Once it loads it fully in memory, it'll have the cost. That was the original fp8 download but its not quite as good quality as the GGUF, despite the speed.
@c0ur4ge Thanks for reply,hope we can have fp8 version in the future, good job.
@zczcg It's up for you :)
What happened to the Q4 models, looks like you took them down and all of them are 11GB now.
Heya! I likely did that a long while back - it was because I was just naively quanting them, running a couple generations and then kicking them out. I couldn't really vouch for quality but I could look at re-upping them or you're free to quant them yourself!
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.