This is a test LoRA trained on 9 images of a specific Azure Lane artwork style.v2.0 was trained on 12 images and kinda improves the quality of the hand.<h3 id="how-to-use:"> How to use:</h3><ul><li>@azlmdia is the style trigger word</li><li>No need to add "high quality, masterpiece" tag.</li><li>For characters not covered in the 12 images, the style is not strong, but there is a subtle difference without using the LoRA.</li><li>Do not increase the model strength; increase the "@azlmdia" tag instead. I recommend setting it from 1.5 to 2.0 for characters outside of Azur Lane. eg "(@azlmdia:2.0)".</li></ul>

Azur Lane Media

Trained on 12 pictures. Hands are now a bit better.

Files with SHA256: 294ee38cecd1c69160049180bc0361255b3075c4fb556f4aa2e26b6c5f456bc6