CivArchive
    "Tareme(タレ目)" set (Eye collection special edition) - Type 2
    Preview 8552343
    Preview 8552348
    Preview 8552350
    Preview 8552347
    Preview 8552346
    Preview 8552353
    Preview 8552358
    Preview 8552357
    Preview 8552355
    Preview 8552354
    Preview 8552351
    Preview 8552352
    Preview 8552356
    Preview 8552349
    Preview 8552344

    (日本語説明は後半にあります)

    Reproduces slanted down eyes, called "Tareme" in Japanese, a familiar anime/comic expression.

    4 shape types are available. To download each types, switch version tabs ↑ above.

    These are special blended editions of my trained data for "Eye collection" LoRA series (document are currently in Japanese only).

    Not suitable for photorealistic models.

    Easy to use. No trigger prompts are necessary.

    More various other SD1/SDXL LoRA's and checkpoints are available on my HuggingFace repository or CivitAI models list. Please check it out.

    CAUTION : Some sample images' prompt is using NegPiP extension for minus weight.


    キャラクターの目つきをタレ目に変更できます。

    形の異なる4種類を用意しました。各タイプをダウンロードするには、ページ上部のバージョンタブを切り替えてください。

    これらのLoRAは、配布中の「アイコレクション」LoRAシリーズ向けの学習データ(アイコレクションとしてはまだリリースされていない新規データも多数含みます)を複数マージし、安定性や再現性を強めた特製バージョンとなっています。

    アニメ系のデータモデル用です。実写系のデータモデルでは検証されていません。

    LoRAを有効にするだけで発動します。階層調整済みですので、基本的には階層指定せず強度指定のみで大丈夫です。

    他にもさまざまなSD1/SDXL用LoRAおよびデータモデルを公開していますので、HuggingFaceリポジトリまたはCivitAIの配布モデルリストをご覧ください。

    掲載しているサンプル画像のプロンプトの一部では、NegPiP拡張機能を使ったマイナス強度指定を使用しています。

    Description

    FAQ

    Comments (7)

    JekSunMar 17, 2024
    CivitAI

    Your models for faces are the best, I use them very often,

    Thank you :)

    Could you make a Lora to look away in different directions, without Lora it is very difficult to do

    JujoHotaru
    Author
    Mar 17, 2024

    Thank you for your interest. I have sometimes tried to change eye direction, but so far I have not had good results.

    crawlingratMar 17, 2024· 1 reaction
    CivitAI

    Do you think you will ever train these for SDXL?

    JujoHotaru
    Author
    Mar 17, 2024· 2 reactions

    Thank you for your interest. Currently I'm focusing SD1.5, but eventually I would like to try XL.

    crawlingratMar 17, 2024· 2 reactions

    @JujoHotaru I would suggest Pony as a Base for training when you do! The LoRA trainer on this site has the option. ;)

    Yui9987Mar 24, 2024
    CivitAI

    Can i have your dataset and train settings? Your lora performance is GREAT!

    JujoHotaru
    Author
    Mar 24, 2024

    Thank you for your interest. Sorry currently source images are confidential, but not so special. All source images are drawn by myself. And, my LoRA training method is called "Copy Machine method". It can be trained easily by using "TrainTrain" WebUI extension's "Difference" method. https://github.com/hako-mikan/sd-webui-traintrain/

    LORA
    SD 1.5

    Details

    Downloads
    576
    Platform
    CivitAI
    Platform Status
    Available
    Created
    3/16/2024
    Updated
    5/12/2026
    Deleted
    -

    Available On (2 platforms)

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