<h3 id="v2.0-description">V2.0 Description</h3><span style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Helvetica, Arial, sans-serif, &quot;Apple Color Emoji&quot;, &quot;Segoe UI Emoji&quot;;font-size:16px">Lora Strength: 0.7 - 1.0This is a LoRA that was created quite some time ago but had not been uploaded until now.V1.0 was experimentally created using Illustrious XL 2.0, but V2.0 has been completely recreated using Illustrious XL 1.0.This is because I believe Illustrious XL 1.0 is superior as a base model for training purposes.だいぶ前に作成したものの、投稿していなかったLoRAです。V1.0では試験的にIllustrious XL 2.0で作成しましたが、V2.0ではIllustrious XL 1.0で新たに作成し直してあります。学習用のベースモデルとしては、Illustrious XL 1.0の方が優れていると考えているためです。<h3 id="trigger-words">Trigger Words</h3>Character:<ul><li>yokoww, 1girl, petite, short stature, green eyes, tsurime, brown hair, short hair, asymmetrical bangs, hair between eyes, v-shaped eyebrows, flat chest, </li></ul>Note: The chest is shown slightly larger in the sample images.※サンプル画像では胸を少し盛ってあります。<h3 id="v1.0-description">V1.0 Description</h3>Created with PixAI <a target="_blank" rel="ugc" href="https://pixai.art/">https://pixai.art/</a>Lora Strength: 0.9 - 1.8Since PixAI now supports training with Illustrious XL v2.0, I created this LoRA as a test.Increasing the weight brings it closer to the original style, but it may also cause unintended side effects.The sample images were generated with a weight of 1.5.PixAIではIllustrious XL v2.0での学習が可能になったため、そのテストとしてLoRAを作成しました。ウェイトを高くすると原作の雰囲気に近づきますが、副作用が出る可能性もあります。サンプル画像はWeight 1.5で作成しました。<h3 id="trigger-words">Trigger Words</h3><h3 id="character:">Character:</h3><pre><code>yokoww, 1girl, petite, slim, green eyes, tsurime, brown hair, short hair, asymmetrical bangs, hair between eyes, v-shaped eyebrows, flat chest, </code></pre><h3 id="outfits:">Outfits:</h3>The model does not include specific training for outfits, so feel free to customize them to your preference.衣装は特に学習させていない為、お好みの衣装を使用してください。(下記はサンプルに使用した衣装)<h3 id="prompt-used-for-the-sample-images">Prompt used for the sample images</h3><ul><li>sportswear, gym uniform, gym shirt, collared shirt, white shirt, short sleeves, gym shorts, white shorts, white socks, sneakers,</li><li>casual, sleeveless shirt, collared shirt, shorts, short shorts, socks, sneakers,</li><li>breasts, schoolgirl underwear, white sports bra, white panties,</li><li>swimsuit, bikini, red bikini, breasts, navel,</li></ul>Since this LoRA was trained on a dataset of monochrome images, it is recommended to add "monochrome, grayscale" to the negative prompt when generating color illustrations.Additionally, including prompts such as "anime coloring" and specifying hair and outfit colors can increase the likelihood of generating color illustrations.For example, try prompts like "brown eyes" or "blue skirt" to clearly indicate colors other than black or white.If the character's trigger word already includes color information, you may not need to worry about this as much.このLoRAはモノクロ画像のデータセットで学習されているため、カラーイラストを生成する際にはネガティブプロンプトに "monochrome, grayscale" を追加することを推奨します。さらに、"anime coloring" のようなプロンプトを含めたり、髪や服の色を指定することで、カラーイラストを生成する可能性を高めることができます。例えば、"brown eyes"や"blue skirt"のように、白や黒以外の色を明確に指定するのが効果的です。なお、キャラクターのトリガーワードに既にカラー情報が含まれている場合は、特に気にしなくても構いません。

Files with SHA256: cc867aaa586a4cf298baaf45271a76517758bf08d1806939bf7db74b7bde9493