I retrained this LoRA using Anima as the base model. While the training dataset remained exactly the same, the resulting style has shifted significantly. In my view, it is not quite as faithful to the original character features as the Illustrious-based version, but it has the distinct advantage of almost never making mistakes when drawing fingers.illustrious model：<a target="_blank" rel="ugc" href="https://civarchive.com/models/3019706/shirasaki-tsugumi">白崎つぐみ Shirasaki Tsugumi (大図書館の羊飼い) - v1.0 | Illustrious LoRA | Civitai</a>recommended lora weight: about 0.8recommended base model：<a target="_blank" rel="ugc" href="https://civarchive.com/models/2458426/anima?modelVersionId=2945208">Anima - base-v1.0 | Anima Checkpoint | Civitai</a><a target="_blank" rel="ugc" href="https://civarchive.com/models/2544636/wai-anima">WAI-ANIMA - v1.0(base 1.0) | Anima Checkpoint | Civitai</a>The "Anima base" model is more faithful to the original, though its art style is somewhat stiff; the "WAI Anima" model, by contrast, features a more elaborate style but ends up distorting the original aesthetic.All the dataset, including trigger words, tags, pictures have been uploaded in optional files, so I wont list all trigger words in this description. just download the training <a target="_blank" rel="ugc" href="http://data.zip">data.zip</a><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">trigger word: shirasaki tsugumi

(Anima ver.)白崎つぐみ Shirasaki Tsugumi (大図書館の羊飼い)

training data.zip

char_anima_lora.safetensors

v1.0

based on illustriousXLRecommened weight: about 0.9Recommened base model:WAI-Branch-Rouwei <a target="_blank" rel="ugc" href="https://civarchive.com/models/1400967/wai-branch-rouwei">WAI-Branch-Rouwei - E-pred-1 | Illustrious Checkpoint | Civitai</a>WAI illustrious <a target="_blank" rel="ugc" href="https://tensor.art/models/990778216270553015/WAI-illustrious-v17">WAI-illustrious - v17 - Free AI CHECKPOINT | </a><a target="_blank" rel="ugc" href="http://Tensor.Art">Tensor.Art</a>If you wish to generate images that most accurately reproduce tsugumi's appearance,choose WAI-Branch-Rouwei.If you wish to generate more elaborate and exquisite images，choose WAI illustriousAll the tags and training data have been uploaded, so I won't list them in description.Just download "training data"zip and you'll get all the tags txt.trigger word: shirasaki tsugumiThanks a lot to Gemini, without it I can't accomplish all the training procedures.There are two versions of lora, a 10 epoch and a 8 epoch. The 10 epoch is more rigorous but somewhat overfitted. 8 epoch is more faithful to the original but exhibits a higher degree of randomness.

Files with SHA256: fd0a1ba97abb9c99d65f43242e807e8a4e59d24c72da544c36d3f546b0af528b