<p>This one turned out pretty good.   It's trained on every page in the Invincible comic series and then finetuned on a small subset of them that I thought would fix the issues with the first draft.   The style should just work most of the time, but if it isn't working with your prompt, or you feel it should be a bit stronger just add <code>illustration</code>.   This is a style lora so while you'll see features from the comics characters none of them were tagged in the training data.  You can get some of them by describing them, but that wasn't my intention when making this.</p>

Files with SHA256: 3d2d568c9f46d4b1b7f13c13316ccd85a826f86f8c9361950583db42ad896d33