CivArchive
    LunaR - v1.0 - v1.0
    NSFW
    Preview 131530843
    Preview 131531476
    Preview 131531471
    Preview 131531477
    Preview 131531481
    Preview 131531480
    Preview 131531484
    Preview 131531485
    Preview 131531489
    Preview 131531488
    Preview 131531492
    Preview 131531491
    Preview 131531494
    Preview 131531495
    Preview 131531496
    Preview 131531497
    Preview 131531501
    Preview 131531502
    Preview 131531505
    Preview 131531506

    LunaR is a specialized LoRA designed to enhance [mention the style, e.g., feminine features / cinematic lighting / specific character details]. It focuses on delivering high-quality textures and consistent results for [mention the subject, e.g., portrait photography / stylized characters].

    Key Features:

    • Style: [e.g., Realistic, Semirealistic, Stylized]

    • Best used for: [e.g., Close-up portraits, dramatic lighting, fashion shots]

    • Trigger Word: LunaR


    Usage Tips

    • Recommended Weight: 0.6 - 0.8 (Adjust based on your checkpoint)

    • Sampling Method: [e.g., Euler a, DPM++ 2M Karras]

    • CFG Scale: 5 - 7

    • Base Model: Recommended for use with [mention your base model, e.g., Pony Diffusion V6 / SDXL 1.0].


    Notes

    • What it’s good at: [e.g., Capturing intricate skin details and soft lighting.]

    • What to avoid: [e.g., Not intended for complex landscapes or heavy architectural work.]

    If you enjoy using this model, please consider leaving a review and sharing your generations! Your feedback helps me improve future iterations.

    Description

    Version 1.0 Release Notes

    • Base Model: Built on the HiDream architecture, optimized for high-fidelity character generation.

    • Core Improvements: * Enhanced photorealism and skin texture detailing compared to standard base models.

      • Improved consistency for the character's facial features and specific accessories (e.g., glasses, nose ring).

      • Better handling of urban fashion aesthetics and varied lighting environments.

    • Visual Style: Focused on a sharp, cinematic look with a refined color palette, suitable for both portraiture and lifestyle-oriented generation.

    • Training Focus: The model has been fine-tuned to maintain character identity across diverse poses and backgrounds, ensuring high-quality output while leveraging the strengths of the HiDream base.

    LORA
    HiDream

    Details

    Downloads
    6
    Platform
    CivitAI
    Platform Status
    Available
    Created
    5/23/2026
    Updated
    5/23/2026
    Deleted
    -
    Trigger Words:
    LunaR

    Files

    clip_l.safetensors

    Size:
    234.74 MB
    SHA256:

    Mirrors

    LUNALORA.zip

    Size:
    206.26 MB
    SHA256:

    Mirrors

    image_z_image_turbo (1) (1) (1) (1).json

    Size:
    6.06 KB
    SHA256:

    Mirrors

    decoder.pt

    Size:
    2.32 GB
    SHA256:

    Mirrors