<h1 id="lora-weights-of-stable-diffusion-xl-for-fast-image-generation.-5k5plrqxq">LoRA weights of Stable Diffusion XL for fast image generation.</h1><p></p><h2 id="important-usage-guidance-kcaoi0u1a">Important Usage Guidance</h2><ol><li><p>Use DDIM or Euler instead of LCM for sampling!</p></li><li><p>The name of each LoRA weights indicates how many inference steps they should be applied.</p></li><li><p>The name of each LoRA weights indicates whether they are able to use normal CFGs or small CFGs.</p><ul><li><p>NormalCFG means that model equipped with the LoRA can use CFG value 2-9 for generation. For example, <strong>Juggernaut XL, </strong>in my test, works with 3-5. Hopefully I just find a better way to train normalCFG LoRAs with better visual quality, will update them in the coming days!!!</p></li><li><p>SmallCFG means that the model equipped with the LoRA can use CFG value 1-2 for generation.</p></li></ul></li></ol><p>About the performance of normal CFG LoRAs.</p><p></p><h2 id="important-information-zh0npnu84">Important Information</h2><p>[[paper](<a target="_blank" rel="ugc" href="https://arxiv.org/abs/2405.18407">https://arxiv.org/abs/2405.18407</a>)]</p><p>[[code](<a target="_blank" rel="ugc" href="https://github.com/G-U-N/Phased-Consistency-Model">https://github.com/G-U-N/Phased-Consistency-Model</a>)]</p><p>[[huggingface](<a target="_blank" rel="ugc" href="https://huggingface.co/wangfuyun">https://huggingface.co/wangfuyun</a>)]</p><p></p>

Files with SHA256: 82decf2091d19ef988dee1a46bb65ead0b8ca1b97009468bc6357b59370b8bba