CivArchive
    【NoobAI-XL】NoobAI-Pastel illustrationXL - v1.0
    NSFW
    Preview 43114166
    Preview 43114027
    Preview 43114169
    Preview 43113910
    Preview 43114063
    Preview 43114069
    Preview 43114048
    Preview 43114028
    Preview 43114057
    Preview 43114029
    Preview 43114064
    Preview 43114040
    Preview 43114044
    Preview 43114066
    Preview 43114068
    Preview 43114020
    Preview 43114184
    Preview 43114128
    Preview 43114126
    Preview 43114129

    The user has made significant improvements to their first-generation traditional Chinese landscape painting LoRA model, focusing on enhancing lighting effects and character expressions. Below is a detailed breakdown of the optimized workflow:

    1. Base Model Recommendations

    • Primary Options: SDXL, NoobAI-XL (NAI-XL).

    • These models ensure compatibility with the enhanced stylistic details and broader tonal range.

    2. Sampling Configuration

    • Method: Euler a.

    • CFG Scale: Suggested range of 3–7, balancing creativity and adherence to prompts.

    3. Resolution Guidelines

    • Preferred output resolutions include:

      • 576x864 (for smaller, detailed compositions).

      • 768x1024 (medium standard for high-quality works).

      • 1024x1536 (for larger, visually impactful scenes).

    4. Upscaling and Redrawing

    • Upscaling Algorithm: R-ESRGAN 4x+ Anime6B.

    • Redraw Strength: Set between 0.25 and 0.5 to retain intricate details while refining edges.

    This refined workflow ensures higher fidelity in artistic expression, particularly in the interplay of light and character features.

    Description

    FAQ

    LORA
    SDXL 1.0

    Details

    Downloads
    267
    Platform
    CivitAI
    Platform Status
    Available
    Created
    12/1/2024
    Updated
    5/14/2026
    Deleted
    -
    Trigger Words:
    Pastel illustration

    Files

    NoobAI-Pastel illustrationXL.safetensors

    Available On (1 platform)

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