<h1 id="splatoonxl-2qubkv4fn">SplatoonXL</h1>Generates colourful squids and octos.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/2ebda6be-e6a3-4c4f-a661-f8642fa46b41/width=525/2ebda6be-e6a3-4c4f-a661-f8642fa46b41.jpeg" /><h2 id="usage-notes-rfe876q40">Usage notes</h2><ul><li>This model will help SDXL-based checkpoints draw:<ul><li>in a style reminiscent of Splatoon concept and promotional 2D artwork.</li><li>concepts from Splatoon, including Inklings and Octolings.</li></ul></li><li>Recommended checkpoints:<ul><li>v3.0: Use a Pony V6 checkpoint, such as Pony Diffusion V6 XL, or AutismMix.</li><li>v2.0/v1.0: Use an anime checkpoint, like kohakuXL, blue_pencilXL, CounterfeitXL, or Yamer's Anime/Unstable Diffusers.</li></ul></li><li>Guidance:<ul><li>Prompting for keywords relating to place, ambience, or an activity can help with drawing scenery around characters.</li><li>Negative prompting is needed to reduce limb deformities and unwanted clutter.</li><li>Use your favourite prompting patterns or these suggestions.<ul><li>v3.0: Prefix positive prompt with "score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up" following Pony guidelines.</li><li>v3.0/v2.0/v1.0: I found "photorealistic, 3d model, bad, worse, worst, ugly, bad anatomy, blurry, close-up, disembodied limb" in the negative prompt (taken from kohakuXL page) to consistently give cleaner images.</li></ul></li></ul></li><li>Have fun!</li></ul><h2 id="release-notes-ty1vmfe7n">Release notes</h2>v3.0 (Pony V6)<ul><li>Retrained against the Pony Diffusion V6 XL checkpoint.</li><li>Training dataset is identical to v2.0.</li><li>Adjusted training parameters.</li></ul>v2.0 (SDXL 1.0)<ul><li>Primary focus of changes were to improve generated Octoling accuracy, and general control over Inkling/Octoling features.</li><li>Cleaned up tagging.</li><li>Added more images to dataset.</li><li>Tuned training parameters.</li></ul>v1.0 (SDXL 1.0)<ul><li>Initial release.</li></ul><h2 id="model-information-6orwlb7mj">Model information</h2>This model was trained against official Nintendo Splatoon artwork. Art was selected by hand to match a particular visual style (highly saturated flat-shaded art).See the "About this version" card for information about the dataset and training parameters. Images, tags, and training configuration are provided in an optional zip file along the model.<h2 id="attribution-r2mzax7pl">Attribution</h2>This work was inspired by the following linked model by IWannaGoBack, which targets a similar niche for SD 1.5 based networks.<a target="_blank" rel="ugc" href="https://civarchive.com/models/85425/splatoon-style">https://civarchive.com/models/85425/splatoon-style</a>Thanks to IWannaGoBack. I enjoyed playing with your model, and it inspired me to dive into the deep rabbit hole of LoRA training.<h2 id="public-domain-notice-and-disclaimers-oovn5ek13">Public domain notice and disclaimers</h2>The author of this generative AI model (model) and curator of its training database (database) releases rights and disclaims liabilities that pertain to the model/database themselves or usage thereof, under the mark of the CC0 1.0 legal code.<a target="_blank" rel="ugc" href="http://creativecommons.org/publicdomain/zero/1.0">http://creativecommons.org/publicdomain/zero/1.0</a>This is a fan work that derives from intellectual property owned by Nintendo Co., Ltd. It is not endorsed by Nintendo.

SplatoonXL

Training run 12, epoch 000008Dataset 8Images<ul><li>Count: 58 images</li><li>Sources: Danbooru (search terms: splatoon official_artwork), Inkipedia, Splatoon Base.</li><li>Treatment: Some images were cropped to isolate individuals. No image modifications beyond cropping.</li><li>Flip augmentation: enabled</li></ul>Tagging<ul><li>Methodology: Tags were scraped alongside images from Danbooru. Images from other sources were tagged by WD14 (wd-v1-4-swinv2-tagger-v2). Tags were briefly reviewed by model author for correctness and consistency.</li><li>Caption shuffling: All tags are randomly shuffled. No trigger words were used.</li></ul>Training<ul><li>Dataset replicates/epoch: 10</li><li>Epochs: 8</li><li>Batch size: 4</li><li>Checkpoint: sd_xl_base_1.0.safetensors</li><li>Network configuration: LyCORIS/LoCon (dim=16, alpha=8, conv_dim=8, conv_alpha=1)</li><li>LR: 0.0002 for both U-Net and text encoder. Algorithm was "cosine_with_restarts" with 5% warmup.</li><li>Optimizer: AdamW</li></ul>

Files with SHA256: 69ed54cf88b9268502305d7f1451532398cce2a028b9d60c5dd0552294f4ff52