CivArchive
    Preview 35695887
    Preview 35696287
    Preview 35695902
    Preview 35696305
    Preview 35696821
    Preview 35696460
    Preview 35696806
    Preview 35696954
    Preview 35696762
    Preview 35697070
    Preview 35697058
    Preview 35698376
    Preview 35698575

    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

    C4PACITOR ANALOG is the first taste of v2. Born from training Flux.1 D on a subset of our UHD analog/digital photos of people and some light NSFW. It is a softer, fun-to-use-with LORA's take on dev trained in full BF16 on 1024x1024 (or higher) images and some other DR34MT34M goodness.

    FAQ

    Comments (12)

    NarzOct 21, 2024· 24 reactions
    CivitAI

    v2 looks like a huge step up.. schnell v2 coming..? 😬

    c0ur4ge
    Author
    Oct 21, 2024

    Thanks! I hope so but I've been lucky enough to keep fresh batches of V2 tested and cycling. Eventually yes, but the fact that its tough to do a full tune on schnell (either) without toasting things makes it tougher; I don't just want to kick out some bad merge so i've got some testing ahead!

    c0ur4ge
    Author
    Nov 25, 2024

    First taste of this is up! Let me know how it treats you! :)

    nedixOct 28, 2024· 13 reactions
    CivitAI

    A solid checkpoint but a bit too many artifacts in image generation (fingers, arms etc.)

    c0ur4ge
    Author
    Oct 29, 2024· 1 reaction

    If you can - please share a prompt example (or dm me) so I can help see if there’s something specific in the dataset that would be the cause.

    That said - I’d also suggest increasing guidance and trying a standard resolution if it’s very narrow/wide.

    nedixOct 29, 2024· 1 reaction

    I DM'ed you a Dropbox link

    roberto_baggioNov 4, 2024· 2 reactions

    @c0ur4ge, Many people have spread the idea that a low guidance scale is a good thing without explaining the drawbacks. Also, We can find many publicly shared workflows with very low guidance in it.
    Now, lots are using very low guidance with everything and wonder why they're having problems.
    Thanks for suggesting increasing guidance, many are not aware of that. ¯\_(ツ)_/¯

    c0ur4ge
    Author
    Dec 18, 2024

    Curious to know how you feel its been faring lately! :)

    Legend_of_Chavo_GNov 3, 2024· 23 reactions
    CivitAI

    Since your taking questions (whether u like it or not) would it help if I requested this for on-site generation? Or did u elect not to have it except eternally?

    c0ur4ge
    Author
    Nov 4, 2024· 1 reaction

    Nope - I've not yet had anyone ask about it but I think that'd be cool! Please feel free to do so! Thank you!

    798thtgNov 3, 2024· 24 reactions
    CivitAI

    Thanks for an update here. Can you confirm that the latest Schnell is fully Apache 2 and you didn't integrate Dev into it?I intend to use it for commercial purposes.

    c0ur4ge
    Author
    Nov 4, 2024· 1 reaction

    Schnell base with only some LORA influence/merge. No dev layers were merged into the 's' checkpoints, no.

    Checkpoint
    Flux.1 D

    Details

    Downloads
    907
    Platform
    CivitAI
    Platform Status
    Available
    Created
    10/21/2024
    Updated
    5/14/2026
    Deleted
    -

    Files

    c4pacitor_aV2Alpha.safetensors

    Mirrors

    Available On (1 platform)

    Same model published on other platforms. May have additional downloads or version variants.