CivArchive
    Preview 6399557
    Preview 6399542
    Preview 6399541
    Preview 6200243
    Preview 6399547
    Preview 6399545
    Preview 6399549
    Preview 6399552
    Preview 6399555
    Preview 6399558
    Preview 6399556
    Preview 6399559
    Preview 6399560

    🐸 Try Stoked Reality XL v1.1 for FREE on Tensor.art 🐸

    πŸ„ Try Stoked Tensor XL for FREE on Tensor.art πŸ„

    πŸ›Ή Try RealFeel XL for FREE on Tensor.art πŸ›Ή

    β˜• If you like what I do, buy Stoke a brew β˜•

    Stoked Reality

    Not released any SD1.5 assets in a while, so gave Stoked Reality a make over. Generates better at higher resolutions than it predecessors, best results with a low CFG. Better eyes and skin with this version also.

    SHOW YOUR APPRECIATION

    To show your appreciation, give a ❀️ to your favorite images, ensuring you receive future updates and news. Join our artistic community and showcase your own original artworks and the masterfully crafted prompts that created them.

    Description

    FAQ

    Comments (3)

    PotravkaFeb 3, 2024Β· 3 reactions
    CivitAI

    StokedReality V3 - this is my new favourite. it does well on 768x1024
    Weird thing happened. I've downloaded this model to test, because long time have no tried anything new from 1.5. Then I accidentally put my old prompt with my custom lora in it (that was trained based on realistic vision long ago) and they worked ugly on realistic vision.. But it works Excellent on this model just as i hoped it should work. It has inhaled new life in my old loras. And it also does great in img2img with teeth and eyelashes on portraits when denoising strength is too high,

    Stoke
    Author
    Feb 4, 2024Β· 1 reaction

    Thank you for your feedback. TBH I hadn't used or done anything with SD1.5 for a while, then realised the amount of people still using it due to hardware limitations. Also a lot of video uses 1.5, which I am just beginning to get into. So thought it was time to start giving 1.5 some love again.
    As this is Version 3.0 and a new beginning I think I will make training, inpainting, fp32 models.
    I had tried it on 768x1152 with it working well the majority of the time, but proportions were beginning to go - especially facially. Glad to hear about LoRA integration as TBH I was pushed for time (end of my 3 weekly days off work) hence the 'One Button Prompt' images ( which when guided a little can give great results) and lack of thorough LoRA testing. There is also the bonus of everything being so much quicker working with 1.5.

    Dennuz666Feb 13, 2024Β· 2 reactions

    I can confirm this. After reading your comment i tested this on different checkpoints and yes ... great results here too, while other checkpoints are failing. Could be my n00b lora creation skills too, but this far pretty awesome.

    Checkpoint
    SD 1.5

    Details

    Downloads
    1,293
    Platform
    CivitAI
    Platform Status
    Available
    Created
    2/3/2024
    Updated
    5/13/2026
    Deleted
    -

    Files

    stokedReality_v30.safetensors

    Mirrors

    stokedReality_v30.safetensors

    Mirrors

    HuggingFace (1 mirrors)

    Available On (1 platform)

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