I don't know why, but the prompts in the image metadata on this LoRA's posts are all mixed up, so please don't rely on them too much xDFor some reason, putting "masterpiece" in the negative prompt actually makes the results look closer to the original training images. But the results actually look better without "masterpiece" in the negative prompt. <pre><code>lora strength: 1.0</code></pre><pre><code>trigger word: franzstyle</code></pre><pre><code>The sample images were created using the er_sde sampler.</code></pre>

Files with SHA256: d584aed5f6a1185481e48aef0260da153c9481263fd1e19c2a98b9136d33f2f2