<h3 id="version-2:">Version 2:</h3>IMPORTANT: The trigger word has changed! <code>kxbrie woman</code>A bigger dataset, with:<ul><li>bigger and more consistently large bust</li><li>better labeling</li><li>more steps in training</li><li>cleaned up dataset, consisting of new images generated from lora v1 including some topless images, and artifacts assiduously removed</li><li>a trigger word that doesn't include numbers or underscores</li></ul>The good:<ul><li>No more jpeg artifacts; images are clean and crisp</li><li>bust is consistently large (though not consistently the same size; remains sensitive to prompting)</li><li>the trigger word isn't printed all over the place</li><li>still good adherence to poses</li></ul>The bad:<ul><li>You might get some nipples you didn't ask for, I can think of a couple places that labeling might have led to this. I'm already considering a v2.1. But this is only happening to me about 1 in 5 images or so.</li><li>If you start to specify anything about her bust, it also grow a whole lot</li><li>It's very picky about what her skin looks like (e.g.: asking for a tan does not work, and the sample photo with the yellow outfit should have ink on her face). I'm sure I can weaken this as well.</li><li>The likeness tends to apply to the image and bleed into other people generated in the same image (this was probably true of v1 as well).</li></ul><h3 id="version-1">Version 1</h3>This model's data set is smaller and lower quality than Chloe (my first character LoRA). I've noticed some artifacts appearing in this character's images that she does not suffer from. I also suspect using an underscore in the trigger word <code>kx_br13 woman</code> wasn't the best idea, as it tends to print it everywhere.<h3 id="about-brie">About Brie</h3>A character for Krea 2 created from a generated dataset.Consistent dark rimmed glasses are present in almost every training photo. If you prompt for glasses, expect them to look similar.

Brie - Accidentally Sexy Blonde

A bigger dataset, with:<ul><li>bigger and more consistently large bust</li><li>better labeling</li><li>more steps in training</li><li>cleaned up dataset, consisting of new images generated from lora v1 including some topless images, and artifacts assiduously removed</li><li>a trigger word that doesn't include numbers or underscores</li></ul>The good:<ul><li>No more jpeg artifacts; images are clean and crisp</li><li>bust is consistently large (though remains sensitive to prompting)</li><li>the trigger word isn't printed all over the place</li><li>still good adherence to poses</li></ul>The bad:<ul><li>You might get some nipples you didn't ask for, I can think of a couple places that labeling might have led to this. I'm already considering a v2.1. But this is only happening to me about 1 in 5 images or so.</li><li>If you start to specify anything about her bust, it also grow a whole lot - assuming you're using a bypass</li></ul>

Files with SHA256: 874f5b57dc21062e1b4fb8e8123944595e745dc542b1a7753f82827d8a66afc9