Today, I bring you Uesugi Kenshin from the Rance series of H-games. The lora is trained using two different datasets to provide two different versions, one for her Sengoku Rance appearance, and one for her Rance X appearance. Since having both of them in the same version would cause conflict with tagging, I instead opted to keep them separate so people can choose the appearance they want.Sengoku Rance version tags:<pre><code>Trigger word: srKenshin</code></pre><pre><code>Appearance: black hair, blue hair, long hair, blunt bangs, blue eyes, hair ribbon, medium breasts</code></pre><pre><code>Outfit: armor, armored dress, japanese armor, shoulder armor, sode, kusazuri, gloves, black socks, black shoes</code></pre><pre><code>Helmet: helmet, kabuto /(helmet/)</code></pre>Rance X version tags:<pre><code>Trigger word: rxKenshin</code></pre><pre><code>Appearance: black hair, blue hair, long hair, blunt bangs, blue eyes, hair ribbon, white headband, medium breasts</code></pre><pre><code>Outfit: armor, armored dress, japanese armor, shoulder armor, sode, kusazuri, gloves, black socks, metal boots</code></pre>As usual, I recommend you use the model at 0.8 strength, or 0.6 for the anima version. Rance himself may appear if you include any males in the prompt, but it shouldn't be that hard to prompt him out, especially if you use this in conjuction with other loras.

Uesugi Kenshin (Rance series) [Illustrious/Anima]

Anima version for her original Sengoku Rance design. Settled for epoch 3 for this one since epoch 5 ended up overtrained. Trust me, you don't want me to upload epoch 5. I also added separated her hair ribbons into two separate tags because her helmet interfered in training for the one at the top of her head. Recommended weight is 0.7, but it should ve very capable of doing well at higher if you prompt well.

Files with SHA256: 362ef735ef642fdf3e18f1b933648b3f2a1dc7a98b32e7b11aeab3e8c8dbcc52