<h1 id="version-notes"><span style="color:rgb(250, 82, 82)">Version notes</span></h1><h3 id="v1.0">v1.0</h3><ul><li><p>First version that is on par/better than the illustrious mix</p></li></ul><hr /><h1 id="model-focus"><span style="color:rgb(250, 82, 82)">Model focus</span></h1><ul><li><p>3DCG/Semi realistic face; sharper/mature facial features with subtle skin texture.</p></li></ul><h3 id="weakness">Weakness</h3><ul><li><p>Noise pattern on image. Recommend using WAN 2.1 fp32 VAE and film grain postprocessing node.</p></li><li><p>Hands can be funky sometimes.</p></li><li><p>Anime style hairstyle can be hard to prompt</p></li></ul><hr /><h1 id="recommended-settings"><span style="color:rgb(130, 201, 30)">Recommended settings</span></h1><h3 id="(can-see-workflow-via-uploaded-images)"><span style="color:rgb(130, 201, 30)"><strong>(can see workflow via uploaded images)</strong></span></h3><h3></h3><p><strong>euler + simple</strong></p><ul><li><p>steps 8-10, cfg 1</p></li><li><p>standard, simple, stable, works most of the time</p></li><li><p>downside - lacks texture, uniform texture, softer looking</p></li></ul><p></p><p><strong>euler + beta/(beta57-more tex)</strong></p><ul><li><p>steps 8-12, cfg 1</p></li><li><p>more texture than simple but can cause more artifacts</p></li></ul><p></p><p><strong>euler_a + beta57</strong></p><ul><li><p>steps 8-15, cfg 1</p></li><li><p>even more texture and details but also can get the image wrong</p></li><li><p>can cause the image to look more illustrative/less realistic</p></li></ul><p></p><p><strong>deis_2m / deis_3m + beta57</strong></p><ul><li><p>steps 3-4, cfg 1</p></li><li><p>similar to euler+beta57 but can have more stable structure while having as much texture</p></li><li><p>might have more realistic proportions compared to euler</p></li><li><p>downside - can be soft?</p></li><li><p>deis_3m is longer but might have better structure</p></li></ul><hr /><h1 id="remarks"><span style="color:rgb(250, 82, 82)">Remarks</span></h1><ul><li><p>Personally use 1024x1536 resolution most of the time. Lower resolutions makes the noise pattern more noticeable when upscaled.</p></li><li><p>Upscale model recommendation: 2xLiveActionV1_SPAN. Keeps the image similar to the original without extra sharpening or lost of details while 2x resolution. Also fast.</p></li><li><p>WAN 2.1 fp32 VAE and film grain postprocessing node for noise pattern reduction.</p></li></ul><p></p>

Files with SHA256: 8aecc4bcaa13b1fcb844470505b24b5bfc90b9f892e7b52c346e64c67996adff