CivArchive
    Realistic Portrait - Midjourney mimic - v1.1 (RealVis5.1)
    Preview 13912510
    Preview 13912500
    Preview 13912519
    Preview 13912516
    Preview 13912506
    Preview 13912504
    Preview 13912501
    Preview 13912505
    Preview 13912513
    Preview 13912518

    Realistic Portrait Style

    Lora_Scale: 0.5

    DPM++ 2M SDE Automatic

    CFG Scale 3,5 - 7 (recommend 4)

    Only positive prompts are needed to generate good results

    Description

    Realistic Portrait Style

    Lora_Scale: 0.4 - 0.8

    DPM++ 2M SDE Automatic

    CFG Scale 3,5 - 7 (recommend 4)

    Recommended with Realvis5.1

    Only positive prompts are needed to generate good results

    FAQ

    Comments (4)

    trongg
    Author
    May 29, 2024· 2 reactions
    CivitAI

    with v1.1 recommend use with SG161222/Realistic_Vision_V5.1_noVAE in huggingface, i find that the checkpoint on civitai doesn't give as good results

    TooaiArtMay 29, 2024

    But this Lora is ofr SDXL or SD15? Realistic_Vision_V5.1 is sd15

    trongg
    Author
    May 29, 2024· 1 reaction

    @TooaiArt yes, v1.0 base sdxl, v1.1 base sd15

    HoCatAug 21, 2024· 2 reactions
    CivitAI

    Using ComfyUI with the SDXL efficiency node, the base model is sd_xl_base_1.0, the Latent Size is set to 1024x1024, the VAE is the one provided by the base model, the sampler_name is dpmpp_2m_sde, the scheduler is karras, steps are set to 50, and CFG is set to 4.0. However, using the prompt information from the second image does not produce the same or a similar result.

    LORA
    SD 1.5

    Details

    Downloads
    757
    Platform
    CivitAI
    Platform Status
    Available
    Created
    5/29/2024
    Updated
    5/21/2026
    Deleted
    -

    Files

    style_lora_r51_update_out.safetensors