CivArchive
    Anything-v3-inpainting - 1
    Preview 22912
    Preview 22917
    Preview 22916
    Preview 22915
    Preview 22914
    Preview 22913

    This is a merge of the "Anything-v3" and "sd-1.5-inpainting" models with the "Add difference" option. This model allows you to do high-quality inpainting in anime style

    Description

    FAQ

    Comments (12)

    poisenberyDec 30, 2022· 3 reactions
    CivitAI

    What exactly is the point of this? Inpainting was already possible with the base model [I do it all of the time].

    wavymulderDec 31, 2022· 12 reactions

    The user posted a good demo on reddit comparing using the base model to inpaint vs using the merge, you can check out the post here: https://www.reddit.com/r/sdforall/comments/zyieht/how_to_turn_any_model_into_an_inpainting_model/

    wavymulderDec 31, 2022· 6 reactions

    From what I understand, the inpainting model actually considers what is in the masked region whereas a normal model will just not touch it, so you can still get good results but with a much lower success rate.

    ShiniechoFeb 13, 2023

    @wavymulder exactly, the original model acts really bad at the borders of the mask, creating obvious seams that hard to fix, this model just nails it.

    daemonMar 14, 2023

    A full model renders the entire image, and then essentially pastes just the masked area - this results in obvious seams, as the masked region blended with the newly rendered whole image.

    An inpainting model only renders the mask, allowing it to actually blend with the surrounding original image properly.

    jejeslebewJan 12, 2023· 5 reactions
    CivitAI
    sd/stable-diffusion-webui/modules/sd_models.py", line 350, in load_model load_model_weights(sd_model, checkpoint_info) File "/content/gdrive/MyDrive/sd/stable-diffusion-webui/modules/sd_models.py", line 207, in load_model_weights model.load_state_dict(sd, strict=False) File "/usr/local/lib/python3.8/dist-packages/torch/nn/modules/module.py", line 1671, in load_state_dict raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format( RuntimeError: Error(s) in loading state_dict for LatentDiffusion: size mismatch for model.diffusion_model.input_blocks.0.0.weight: copying a param with shape torch.Size([320, 9, 3, 3]) from checkpoint, the shape in current model is torch.Size([320, 4, 3, 3]). Stable diffusion model failed to load, exiting


    I got this error when loading the model, why?

    MindInTheDigits
    Author
    Jan 13, 2023

    Did you rename the model file? The file name must have "-inpainting" at the end. For example, Any3-inpainting.ckpt

    MintsevJan 26, 2023

    @MindInTheDigits Alas, I have the same problem. Perhaps a .yaml file is needed?

    azerty666Jan 26, 2023

    same problem

    where yaml

    jejeslebewJan 27, 2023

    Thank you. I already fix it by download it without idm. It seems download large model can cause it corrupt. But, now I still got ^C error, it seems like large inpainting model consuming much ram for mine

    zeroxoxoMar 25, 2023

    Had the same error, fixed by using 1.5-inpainting vae instead of anything vae

    bbdbyFeb 24, 2023
    CivitAI

    Any chance of uploading a Safetensor? It would be appreciated!

    Checkpoint
    SD 1.5

    Details

    Downloads
    4,650
    Platform
    CivitAI
    Platform Status
    Available
    Created
    12/29/2022
    Updated
    5/11/2026
    Deleted
    -

    Available On (1 platform)

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