<h2>请视情况选用各版本！</h2><h2 id="please-choose-the-appropriate-version-as-needed">Please choose the appropriate version as needed！</h2>v1.0 preview：这是一个水彩风格的lora。基于anima base v1底模，使用52张图像训练了650步。 它只是一个测试版本。效果可能不稳定。 提示词请参考例图。建议权重1左右。Watercolor Style LoRA - Preview VersionBase Model: Anima Base v1.0 Training: 52 images, 650 steps⚠️ This is a test version. Results may vary.Refer to example images for prompt guidance.Strength: ~1.0 v1.0 base：该版本使用180张图像，两个分辨率训练。这是我第一次尝试多分辨率训练。提示词请参考例图。 和preview版本不同，该版本建议的权重是0.8。例图均为单lora使用0.8权重生成。不过也可以尝试一下1权重。该模型的效果可能不太稳定。This version was trained on 180 images with dual resolution training. This is my first attempt at multi-resolution training. Please refer to the example images for prompts.Unlike the preview version, the recommended weight for this version is 0.8. All example images were generated using this single LoRA at 0.8 weight. However, you can also try weight 1.0. Generation results may be unstable.v1.1 base:该版本使用210张图像，两个分辨率训练。相较于1.0 base，其人物皮肤方面做了改善，现在其皮肤风格更贴近水彩。我认为它总体上不错，不过其细节方面欠佳。建议权重为0.7至0.8，例图的权重均在0.7到0.8之间。<span style="background-color:rgb(21, 21, 23);color:rgb(249, 250, 251);font-family:quote-cjk-patch, Inter, system-ui, -apple-system, BlinkMacSystemFont, &quot;Segoe UI&quot;, Roboto, Oxygen, Ubuntu, Cantarell, &quot;Open Sans&quot;, &quot;Helvetica Neue&quot;, sans-serif;font-size:16px">This version was trained on 210 images at two resolutions. Compared to the 1.0 base, it has improvements in character skin rendering, which now leans more towards a watercolor style. I think it is generally good, though the fine details are somewhat lacking. A strength of 0.7 to 0.8 is recommended, and all example images were generated within that same strength range.

Files with SHA256: b06f78ad3ffe33ab83a6e855fb57bddfc229e2dc820496efd1f4f08c8a8d7165