<h1 id="this-is-a-model-with-a-style-lora-merged-into-the-anima-base-or-animayume">This is a model with a Style LoRA merged into the <a target="_blank" rel="ugc" href="https://huggingface.co/circlestone-labs/Anima">Anima Base</a> or <a target="_blank" rel="ugc" href="https://civarchive.com/models/2385278?modelVersionId=2782261">AnimaYume</a></h1><h1 id="it-may-not-be-suitable-as-a-base-model-for-training.-it-is-recommended-to-use-anima-base-or-animayume-as-the-base-for-training.">It may not be suitable as a base model for training. It is recommended to use Anima Base or AnimaYume as the base for training.</h1><h1 id="this-checkpoint-is-not-an-sdxl-based-model.">※※※ This checkpoint is not an SDXL-based model. ※※※</h1><h1 id="generation-settings">Generation settings</h1><ul><li>resolution</li></ul>1024x1024, 896x1152, 1152x896, etc.1535x1536, 1344x1728, 1728x1344, etc.<ul><li>steps</li></ul>30～50<ul><li>CFG</li></ul>3.0 ～ 5.0<h1 id="prompt">Prompt</h1><ul><li>Quality tag</li></ul><pre><code>masterpiece, best quality, good quality, score_9, score_8, score_7, highly detailed, highres, absurdres, newest</code></pre>Since this model is merged with a style LoRA, the use of @artist —which significantly affect the visual style—is not recommended.<ul><li>Negative tag</li></ul><pre><code>worst quality, low quality, score_3, score_2, score_1, old, oldest</code></pre>Including the number of fingers and toes in the prompt can help improve the results.<ul><li>Positive</li></ul><pre><code>5-fingers, 1-thumb, 5-toes</code></pre><ul><li>Negative</li></ul><pre><code>6-fingers, 6-toes</code></pre><a target="_blank" rel="ugc" href="https://huggingface.co/circlestone-labs/Anima">Please refer to this as well.</a>Version 1.2 has been discontinued.

HarmoniQMix Anima

The training resolution of the LoRAs to be merged has been changed to 1024.

Files with SHA256: be60c4b00a09ac1d1f706c3c8177a325a105f6de5d25a0a701399da09bd25f2f