This is mainly a style lora for <a target="_blank" rel="ugc" href="https://store.steampowered.com/app/1735700/Back_to_the_Dawn/">Back to the Dawn</a> official arts and pixel arts. But it can also do some major characters from game. Back to the Dawn is a prison breaking + crpg genre game, setting in a world full of anthros, where you play as a falsely accused journalist or a undercover agent/cop. It has stylized pixel arts and HD2D scenes.Trained and tested on <a target="_blank" rel="ugc" href="https://civarchive.com/models/833294?modelVersionId=1116447">NoobAI-XL Eps 1.1</a>.Base Prompt format: <code>{art style tag}, {character tag|species}, anthro, solo, {gender tag}, ...</code><h2 id="artwork-style-tags">Artwork Style Tags</h2><ul><li><code>official art</code>: the style of the game's steam cover art.</li><li><code>pxiel \(artwork\)</code>: the escape ending pixel art. If it generates sprite size character/composition, try view point tags such as <code>half-length portrait</code>, or img2img with high noise on a non-pixel generation of good composition.</li></ul><h2 id="other-style-tags">Other Style Tags</h2><ul><li>Single sprite: <code>hd pixels, sprite, white background</code>, generate at 1024×1024.</li><li>Sprite sequence: <code>hd pixels, sprite, sequence, multiple {poses|angles}, white background</code>, generate at 1024×1024. There is randomness in sprite size because different species in the game have different heights.</li><li>In-game 2D scene: <code>hd pixels</code>, this style is weaken by sprites. Better up the lora strength, avoid background tags associated with deep scenery and depict floor/wall/furniture by tags.</li><li>In-game HD2D scene: <code>hd pixels, 3d background</code>.</li><li>Semi-realism: <code>photorealism</code>, learned from a few official fursuit photo. However, noob itself is bad at realism so don't expect much realism. If your subjects is usually depicted in very toony/anime style, the result might be strange. It's suggested to upweight this tag by 1.2 and delay this tag for 10%~20% steps with prompt editing.</li></ul>Here are how each style generalize on characters from other series (legoshi from Beastars for example):<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/ab4ed0cc-0aae-4353-9ac5-65a4be9646d6/width=525/ab4ed0cc-0aae-4353-9ac5-65a4be9646d6.jpeg" /><h2 id="character-tags">Character Tags</h2>There are lots of characters in the game and only a few are brought into the dataset. Our two protagonists (Thomas and Bob) has clear non-pixel official art, while the others only have pixel sprites. The lora can generalize what other trained sprite-only characters look like to some extent.Here are character tags in dataset:<ul><li><code>thomas fox</code></li><li><code>bob black panther</code>, add <code>athletic male</code> if needed<ul><li>causal attire: <code>black jacket, red shirt, jeans</code></li></ul></li></ul><ul><li><code>angelo wolf</code><ul><li>canon attire: <code>trilby \(hat\), white collared shirt, green vest, grey pants, black shoes</code></li></ul></li></ul><ul><li><code>bruce mastiff</code>, add <code>musclegut</code> if needed, wears <code>police uniform, police hat</code></li><li><code>billy bull terrier</code>, wears <code>police uniform, police hat</code></li><li><code>chris lynx</code></li><li><code>crunchy crocodile</code>, add <code>muscular male</code> if needed, he always has his <code>towel around neck</code></li><li><code>fenni wolf</code></li><li><code>maggie white ferret</code></li><li><code>beth poodle</code></li><li><code>john tiger</code>, add <code>muscular male</code> if needed</li><li><code>valdimir polar bear</code>, wears <code>white framed eyewear, semi-rimless eyewear</code>, has <code>arm scar</code></li><li><code>reed red panda</code>, a little too young in dataset, wears <code>glasses</code></li><li><code>whitey rabbit</code></li><li><code>nood raccon</code>, his name is noodle but I have to avoid confusing with the actual food</li><li><code>wesley cheetah</code><ul><li>canon attire: <code>red jacket, purple shirt, rolled up sleeves, grey pants, black boots</code></li></ul></li></ul>For prisoner uniforms in game: <code>{red|orange|blue} prison uniform</code><h2 id="tips">Tips</h2><ul><li>Negatives: <code>worst quality, abstract background</code></li><li>When using <code>pixel \(artwor\)</code>, putting <code>sprite</code> in negative will give smoother shading and less pure colors.</li><li>The pixel arts in training data are mostly "8 physical pixels = 1 logic pixel". Downscale generated image by 8 to get a pixel perfect version.</li></ul>

Files with SHA256: 5eac6d1f1d009b13dc85c0b4b921ea5d5b17308421206858a26b0ecd4735a8d7