CivArchive
    PRMJ - v1
    NSFW
    Preview 71204
    Preview 71216
    Preview 71218
    Preview 71215
    Preview 71214
    Preview 71213
    Preview 71212
    Preview 71211
    Preview 71209
    Preview 71208
    Preview 71207
    Preview 71206
    Preview 71205

    A general purpose model finetuned on a dataset of approx. 28k images (NO NSFW). This is based on 2.1 768 version of stable diffusion. You can also make use of 2.x embeds to enhance your generations.

    Description

    FAQ

    Comments (16)

    ritcher1Feb 3, 2023· 2 reactions
    CivitAI

    Can you share the safer safetensors format ?

    shadowxshinigami
    Author
    Feb 3, 2023

    Added

    ritcher1Feb 3, 2023

    @shadowxshinigami Thank you !

    mousewritesFeb 3, 2023· 6 reactions
    CivitAI

    Hell yeah, we've needed more 2.1 models, epic dude!

    107145Feb 4, 2023· 4 reactions
    CivitAI

    Awesome model. Noticed that you use an "nfixer" embedding. Tried to find it but couldn't, do you have a link?

    rockedtFeb 4, 2023

    I was going to ask the same question. Please link the embedding for negative prompts ^^

    scruffynerfFeb 9, 2023

    After the Discord crowd went hunting for this mystery embed, we eventually learned it'll be released on HF this weekend.

    107145Feb 9, 2023

    @scruffynerf I did eventually find it in a random Discord, but good to hear!

    PotravkaMar 18, 2023

    if i am not mistaken there are illuminati diffusion that has link to nfixer nrealfixer nartfixer

    DeniaudFeb 4, 2023
    CivitAI

    Good work! Could you share the finetune algorithm?

    johnreaderFeb 5, 2023· 1 reaction
    CivitAI

    This is an amazing work. I wonder if you could make this available in Diffuser format on Huggingface? I have some issues loading this into A1111.

    65473Feb 8, 2023
    CivitAI

    looks really nice! please make a diffusers conversion for us pleebs

    noodles3211z_Feb 14, 2023
    CivitAI

    Hey, thanks. can you share how you captioned the training set? I've been trying to fine tune SD 2.1

    shadowxshinigami
    Author
    Feb 14, 2023

    It was a mix of Manual + Blip + Clip

    vvv1Feb 20, 2023
    CivitAI

    Awesome SD 2.1 model!

    One problem is I can't merge it with any other 2.1 768 model in automatic1111 gives this error with any other model merges:

    RuntimeError: [enforce fail at ..\c10\core\impl\alloc_cpu.cpp:72] data. DefaultCPUAllocator: not enough memory: you tried to allocate 41943040000 bytes.

    Have 32GB ram, paging file up to 20GB (+ 3090 GPU)

    Other 2.1 768 models merge fine, just not with prmj

    jamesxueji539May 16, 2023· 1 reaction
    CivitAI

    Hello, could you please provide your training parameters?

    Checkpoint
    SD 2.1 768

    Details

    Downloads
    2,701
    Platform
    CivitAI
    Platform Status
    Available
    Created
    2/3/2023
    Updated
    5/13/2026
    Deleted
    -

    Files

    prmj_v1.safetensors

    Mirrors

    CivitAI (1 mirrors)
    TensorArt (1 mirrors)

    prmj_v1.ckpt

    Mirrors

    CivitAI (1 mirrors)

    prmj_v1.yaml

    Mirrors

    CivitAI (83 mirrors)

    Available On (1 platform)

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