Weight: 0.8 - 1.0
Nyako (Torichamaru)
Trigger Words
nyako, dark skin, dark-skinned female, yellow eyes, blonde hair, multicolored hair, short hair, ahoge, small breasts, 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、お好みのスタイルで。(下記は原作風の衣装)
Original-style outfit
crop top, sleeveless turtleneck, ribbed sweater, bare shoulders, navel, midriff, short shorts, denim shorts, thighhighs, school uniform, jacket, long sleeves, pleated skirt, bowtie, bow, fake animal ears, cat ears, cat tail, black sweater, sleeveless turtleneck, navel, highleg panties, animal hands, paw gloves, paw shoes, cat paws, claws, 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。
Description
FAQ
Comments (2)
Good to see you're okay, you'd been silent so long i was getting worried. I'm running her through the checkpoints and she seems to work pretty well with them. My main problem is getting the existing checkpoints to replicate the original style of the manga/artist. Unfortunately my computer can't use any of the python/github programs and has to rely on website services. The problem is more that Illustrious doesn't have much checkpoint variety or helpful style models, hopefully that side of things improves in the coming months. Putting her in monochrome or greyscale seems to help bring out details for some reason.
I often have other tasks running on my PC, so I can’t always work on processing materials, which makes LoRA creation quite irregular. As for the details, most of the training images are monochrome, so that might be why it reproduces finer details better in grayscale.
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.










