First version of this LoRA, released. All images were generated on CIvitai.
This LoRA was basically trained on some images I've generated in Flux, I thought they looked really good and high quality enough to train them. It has a very retro-ish illustration style, definitely focused more on the blacks and high contrast, but it can be flexible enough, you just need to counter it on the negative, I believe. I haven't tested that out yet.
The first version is literally just the Epoch 10, and as you guys can see through the images, it's looking really good, even with caption prompting. I'll release Epoch 11 aswell, but I do feel like it might be a bit overworked, I don't know, I'll need more feedback on that version.
Recommended:
I recommend using a strength of 0.8~0.9, it gets the job done quite well enough, but you can lower it for more flexibility.
Same as my other LoRAs, here's my detailing prompts:
masterpiece, best quality, ultra-detailed, highest quality, very aesthetic, noxthr0s, very awa, newest, latest
And as for my Negative:
loli, child, kid, young, ****, teenager, children, chibi, moe, camel toe, taut clothes, oversaturated, (watermark, signature:1.3), (worst quality, low quality:1.3), sepia, unfinished, lossy, downscaled, pixelated, lowres, artistic error, fewer digits, extra digits, bad hands, bad feet, jpeg artifacts, old, oldest, patreon logo, watermark, signature, logo, web address, english text, abstract, overweight, obese, chubby, fat rolls, love handle, pudgy, hip bones, hip dip, plump, plump belly, bad anatomy, mutation, deformed, extra limb, missing limb, disembodied limb, disembodied penis, muted color
Thank you again for using my LoRA, you can support me through either donating Buzz, or you can use one of the links provided beneath Profile.