[SD 1.5][Checkpoint]ReV v1.3
<h3><u>Update 2:</u></h3><p>ReV_3 released. Read model version description for more info</p><p></p><h3><u>Update</u></h3><p>ReV_2 released. Read model version description for more info<br /><br /><u>Intro</u></p><p><br />This model is mainly intended for <strong>Portraits</strong> and <strong>Full Body (Depth of Field) </strong>like semi-photorealistic pictures. This <em>can </em>do landscape however, not well. I like this model, jack of all trades, master of none. <strong><u>No restrictions on use of this model, just credit to me and other model creators used in this merge.</u></strong></p><p></p><h3><u>Workflow</u></h3><p>minimal workflow logic based on Protogen's: <a target="_blank" rel="ugc" href="https://www.reddit.com/r/StableDiffusion/comments/1079c0d/protogen_checkpoint_merging_data_reference/"><u>Reddit</u></a></p><p>(Mainly used Difference sum from SD1.5-pruned as tertiary model)<br /></p><h3><br /><u>Recommends</u></h3><p>Resolution:</p><ul><li><p><strong>2:3</strong> ratio (i.e 512x768)</p></li><li><p><strong>1:1</strong> ratio (i.e 512x512 or 576x576)</p></li></ul><p></p><p>Best Samplers:</p><ul><li><p><strong>Euler a <em>(Recommended)</em></strong></p></li><li><p><strong>DPM++ 2M Karras</strong></p></li><li><p><strong>DPM++ SDE Karras</strong></p></li><li><p><strong>PLMS</strong></p></li></ul><p></p><p>CFG scale:</p><ul><li><p><strong>Portraits</strong>: 7-9</p></li><li><p><strong>Landscape</strong>: 9-12</p></li></ul><p></p><p>Steps:</p><ul><li><p><strong>Portraits</strong>: 30-40</p></li><li><p><strong>Landscape</strong>: 40-100</p></li></ul><p></p><p><strong>Hires.fix: </strong></p><ul><li><p><strong>Upscaler:</strong> ESRGAN_4x or ESRGAN_4x+</p></li><li><p><strong>Upscale by: </strong>1.5-2</p></li><li><p><strong>Denoising </strong>strength: 0.3-0.5 (0.3 used in samples)</p></li><li><p><strong>Clip</strong>: 1</p></li></ul><p></p><p></p><p>For best results use the: <a target="_blank" rel="ugc" href="https://huggingface.co/stabilityai/sd-vae-ft-mse-original/tree/main"><u>Standard SD VAE</u></a></p><p></p><p><a target="_blank" rel="ugc" href="https://drive.google.com/file/d/1kJsQ4GSoPDfuworuajvTo4xIS-GJUkkX/view">bad-image-prompt-v2</a></p><p></p><p>SFW: put 'NSFW' and 'uncensored' in negative prompt, SFW in regular prompt</p><p>NSFW: put 'SFW' and 'censored' in negative prompt, NSFW in regular prompt<br /></p><p>Less negative prompts the better</p><h3><br /><u>License</u></h3><p><a target="_blank" rel="ugc" href="https://huggingface.co/spaces/CompVis/stable-diffusion-license"><u>creativeml-openrail-m</u></a></p>
页:
[1]