[SD 1.5][Checkpoint]LOFI LOFI V1 fp16
<h1>L.O.F.I: Limitless Originality Free from Interference</h1><p></p><ul><li><p>? No special face alignment</p></li><li><p>? Improve line details</p></li><li><p>? Improve prompt understanding</p></li></ul><p></p><h3><strong>Prompt suggestions</strong></h3><ul><li><p>Since the text-encoder used is enough trained version, do not use a very high attention to control weight, which will cause some wrong drawing, it is recommended that all attention weights be no higher than 1.2</p></li><li><p>If there is no special composition requirement, there is no need for a lot of negative prompts, such as "missing hands", which may affect human body drawing, and <a target="_blank" rel="ugc" href="https://civitai.com/models/4629/deep-negative-v1x">DeepNegative </a>is enough for negative prompts</p></li><li><p>It is strongly recommended to use hires.fix to generate, Recommended parameters:</p><ul><li><p>(final output 512*768)</p><p>Steps: 20, Sampler: Euler a, CFG scale: 7, Size: 256x384, Denoising strength: 0.75, Hires upscale: 2, Hires steps: 40, Hires upscaler: Latent (bicubic antialiased)</p></li></ul></li></ul><p></p><p>Most of the sample images are generated with hires.fix</p><p><em><u>Note: that if you use hires.fix, you may not be able to reproduce the image with the same set of parameters in webui, because hires.fix introduces double randomness</u></em></p><p></p><h3>Model list</h3><ul><li><p>Grapefruit <a target="_blank" rel="ugc" href="https://civitai.com/models/2583/grapefruit-hentai-model">https://civitai.com/models/2583/grapefruit-hentai-model</a></p></li><li><p>URPM <a target="_blank" rel="ugc" href="https://civitai.com/models/2661/uber-realistic-porn-merge-urpm">https://civitai.com/models/2661/uber-realistic-porn-merge-urpm</a></p></li><li><p>deliberate <a target="_blank" rel="ugc" href="https://civitai.com/models/4823/deliberate">https://civitai.com/models/4823/deliberate</a></p></li><li><p>qgo-10b <a target="_blank" rel="ugc" href="https://civitai.com/models/4188/qgo-10b">https://civitai.com/models/4188/qgo-10b</a></p></li><li><p>HassanBlend <a target="_blank" rel="ugc" href="https://civitai.com/models/1173/hassanblend-1512-and-previous-versions">https://civitai.com/models/1173/hassanblend-1512-and-previous-versions</a></p></li></ul><p></p><p>(To be verified)<s> The weights of any byte in LOFI does not come from dreamlike model.</s></p><p></p><h3>Model merge strategy</h3><p>tldr: unet merge by grapeHentai(text-encoder,1 head layers) + hassanblend(IN-MID layers) + photorealistic(merged, OUT layers)</p><p>There are also some complicated merge modes, so I won’t write them here, and probably no one will read them?</p><p>(Welcome to <a target="_blank" rel="ugc" href="https://discord.gg/v5HFg47J6U">my discord server</a> to discuss if you are interested in model merging)</p><p></p>
页:
[1]