CivArchive
    Makoto Shinkai - Flux - V1
    Preview 26033126Preview 26033692Preview 26033693Preview 26033127Preview 26033125Preview 26033130Preview 26033152Preview 26033149Preview 26033147Preview 26033142Preview 26033148Preview 26033163Preview 26033183Preview 26033186Preview 26033188Preview 26033191Preview 26033497Preview 26033962Preview 26033129

    README.md


    TensorArt : https://tensor.art/u/637417127464347467/models

    🎨 Makoto Shinkai Style - Flux: Hello World! 🌌

    This is a LoRA for capturing the breathtaking style of Mr. Makoto Shinkai! Trained meticulously on a handpicked dataset from Blu-Ray 2K video resolutions, featuring over 500+ stunning landscapes. 🌄

    ⚠️ Pro Tip: Some subtitles might sneak into your renders, so it's recommended to lower the LoRA scale to avoid unwanted artifacts. I found that a scale between 0.1-0.6 works best, with an ideal setting at 0.8 for a balance of quality and style.

    My recommended settings:

    • LoRA Scale: 0.8

    • Steps: 50 (Adjust as needed, but I often use 50 on fal.ai for quick test generations!)

    • CFG: 3.5

    • Trigger Word: None (However, if you'd like to enhance human characters, try adding "Suzume style" to your prompt.)

    🎉 I'd love to see what you create with this model! Please share your masterpieces—it would greatly contribute to its evolution. Don't forget to press the like button if you like or dislike if the model doesn't match what you want (≧▽≦), Thank you and happy creating! ✨

    Thanks to:

    Fal.ai for supa fast generate

    Bulanime from Tensor because she lent his vast.ai account

    Description

    V.1

    For this version is a little less flexible for writing long sentences, sorry!

    LORA
    Flux.1 D

    Details

    Downloads
    253
    Platform
    CivitAI
    Platform Status
    Deleted
    Created
    4/24/2025
    Updated
    5/6/2025
    Deleted
    4/24/2025

    Files

    SujatmikoArafuru_Flux.safetensors

    Mirrors

    763398_training_data.zip

    Mirrors

    CivitAI (1 mirrors)