<h3 id="heading-14">V5.0 update </h3><p>After fourteen versions of iterations, the performance is very good, and simple prompts can also have good effects.   </p><p>Example:</p><p>Watermiku, 1Girl, romanticized realism, solo, long hair, looking at viewer, water,     </p><p>Body edge light, depth of field,&lt;lora: watermikuV14_ SDXL_ Lora-00002:0.7&gt;</p><p><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/e06f3aad-0635-42fb-ab8e-afa4602712ab/width=525/e06f3aad-0635-42fb-ab8e-afa4602712ab.jpeg" /></p><p></p><hr /><p></p><hr /><h3 id="heading-2">V4.0 UPDATE   </h3><p>Added some new datasets and modified some training parameters.   </p><p>Hair has a stronger sense of water quality   </p><p></p><hr /><h3 id="heading-5"></h3><hr /><h3 id="heading-5"></h3><h3 id="heading-5">V3.0 update    </h3><p><span style="color:rgb(193, 194, 197)">More aesthetically pleasing in terms of body transparency, will stop training after updating 1-2 versions   </span></p><p></p><hr /><p></p><hr /><p><span style="color:rgb(193, 194, 197)"> </span></p><h3 id="heading-29">V2.0 update   </h3><p><span style="color:rgb(193, 194, 197)">The new model, which took two days to iterate through five versions, has a stronger artistic expression    </span></p><p></p><hr /><p></p><hr /><p></p><p>Miku's water related aesthetic photos will continue to improve this model.   </p><p>weight recommend:0.6-0.8   </p>

watermiku

watermikuV5_SDXL_lora.safetensors

Files with SHA256: 72cf581490752a2218c533bbaa9ed956ed3bf54e1ac240390f7c9836769ad44f