<h2 id="heading-5664">what is bittensor?</h2><p>Bittensor is pioneering the decentralized production of Artificial Intelligence.</p><h2 id="heading-5665">what is niche image?</h2><p><strong>Niche Image - Subnet 23 - High Quality Images Tailored to Your Need</strong></p><p>This subnet offers large scale inference of machine learning models to be used in production.</p><p>Currently, we provide inference api and synthetic service for image generation above 2000 high quality images / minutes. We're serving many SOTA models: MidJourney, DreamShaperXL, JuggernautXL with tailored features: Sticker Maker, Face To Many, InstantID</p><ul><li><p>Generation Studio: <a target="_blank" rel="ugc" href="https://nicheimage.streamlit.app/">https://nicheimage.streamlit.app/</a></p></li><li><p>Data Synthetic Gallery: <a target="_blank" rel="ugc" href="https://nicheimage.streamlit.app/NicheDataset">https://nicheimage.streamlit.app/NicheDataset</a></p></li></ul><p><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/5bdb7e18-6031-4987-9fed-33491d99a225/width=525/5bdb7e18-6031-4987-9fed-33491d99a225.jpeg" /></p><h2 id="heading-5666">PixelArt XL</h2><h3 id="heading-5667">Descriptions</h3><p>This version of LORA created from MidJourney V6 dataset, belong to NicheImage</p><ul><li><p>Data Synthetic Template</p><ul><li><p>Prompt: "{random_prompt}, Pixel Art --style raw --v {random_ratio} --v 6</p></li><li><p>random_prompt is get randomly from real life image caption: <a target="_blank" rel="ugc" href="https://huggingface.co/datasets/toilaluan/livis-gpt4v-laicon-coco-aes-caption">https://huggingface.co/datasets/toilaluan/livis-gpt4v-laicon-coco-aes-caption</a></p></li></ul></li></ul><p>Niche Image database - <a target="_blank" rel="ugc" href="https://nicheimage.streamlit.app/NicheDataset">https://nicheimage.streamlit.app/NicheDataset</a></p><h3 id="heading-5668">How to use?</h3><p>Recommended parameters</p><ul><li><p>lora_scale = 1</p></li><li><p>prompt template: "{main object}, Pixel Art, highly detailed"</p></li><li><p>negative prompt: "bad_quality, worst quality"</p></li><li><p>base model: JuggernautXL or DreamShaperXL works well</p></li></ul><p>It's the first version of PixelArt XL, we will provide more robust LoRA in the future</p>

Files with SHA256: c9a3bff7c6d4482d26d415dd883a9ea1f497499e82de60c450dc6516d4ddd72f