0xSoul 发表于 2023-6-7 22:58:28

[SD 1.5][Checkpoint]Gllltchy V1

<p>Originally posted to <a href="https://huggingface.co/RandallAware/gllltchy" rel="ugc" target="_blank">HuggingFace by RandallAware</a> and <a href="https://www.reddit.com/r/StableDiffusion/comments/zc3d7p/introducing_my_first_model_gllltchy_a_model_based/" rel="ugc" target="_blank">shared on Reddit</a></p><p>My first attempt at a model, this should probably be called a beta, but since the model is based around a glitchy type style, I figured it goes along with the premise, if the model itself just happens to be a bit glitchy. Please don't download if you want a perfect model, with a stable predictable style. This is not that model. This model is really unpredictable, and different types of results at different resolutions it seems. Definitely for abstract creative types.</p><p>Sometimes it's a bit finnicky and unpredictable, and depending on your prompt you may need to add or subtract weight by using multiple parenthesis like (((Gllltchy))) or (Gllltchy style:0.7). If you're unfamiliar with how to use weights, I'd recommend taking a look at the <a href="https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Features#attentionemphasis" rel="ugc" target="_blank">automatic1111 repo wiki and learning about it</a>.</p>
页: [1]
查看完整版本: [SD 1.5][Checkpoint]Gllltchy V1