Use <a target="_blank" rel="ugc" href="https://e621.net/tags?commit=Search&amp;search%5Bhide_empty%5D=1&amp;search%5Bname_matches%5D=%2A&amp;search%5Border%5D=count">e621 tags</a> (no underscore), Artist tag very effective in YiffyMix.<a target="_blank" rel="ugc" href="https://drive.google.com/drive/folders/14KBOBP1TVNY1DIvbzw2ljAM-pjEpsz3l">GridList Species/Artist(v64) update!! </a>&amp; <a target="_blank" rel="ugc" href="https://mega.nz/folder/UBxDgIyL#K9NJtrWTcvEQtoTl508KiA">LoRAs (SDXL)/samples/wildcards</a>Recommend <a target="_blank" rel="ugc" href="https://mega.nz/folder/AJBxGRQY#Q5Slojcs0RrBHyaiTcUyWA">Artist tags (NoobXL) </a>&amp; <a target="_blank" rel="ugc" href="https://mega.nz/folder/UBxDgIyL#K9NJtrWTcvEQtoTl508KiA/folder/ERhywLpJ">comfyUI workflow</a>.<h2 id="example-setting-(use-recommended-sampler-to-get-correct-quality)">Example Setting (Use recommended sampler to get correct quality)</h2>Setting SDXL (SDXL-lightning &amp; NoobXL)<ul><li>Steps = 12~24</li><li>Sampler = "DDPM", "Euler A SGMUniform", "Euler SGMUniform"</li><li>CFG scale = 3~4</li><li>Negative embeddings SDXL = <a target="_blank" rel="ugc" href="https://civarchive.com/models/148131?modelVersionId=166373">ac-neg1</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/148131?modelVersionId=166375">ac-neg2</a> (You don't really need this)</li><li>Postive LoRA SDXL = <a target="_blank" rel="ugc" href="https://huggingface.co/feffy/seaart-quality/tree/main">SeaArt Quality Tags LoRA </a>(You don't really need this)</li><li>Stop at CLIP layers = 2</li></ul>Setting SD 1.5 + vpred<ul><li>Steps = 30~40</li><li>Sampler = "DPM++ 2M Karras", "DPM++ SDE Karras", "DDIM", "UniPC"</li><li>CFG scale = 6~8</li><li>Negative embeddings = <a target="_blank" rel="ugc" href="https://civarchive.com/models/16807/useful-quality-embeddings-animeillustdiffusion?modelVersionId=53718">deformity_v6</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/84134?type=Training%20Data">bwu </a>[SD-WebUI\embeddings]</li><li>Stop at CLIP layers = 1</li><li>SD VAE = <a target="_blank" rel="ugc" href="https://civarchive.com/models/118561/anything-kl-f8-anime2-vae-ft-mse-840000-ema-pruned-blessed-clearvae-fp16cleaned">kl-f8-anime2,</a><a target="_blank" rel="ugc" href="https://civarchive.com/models/23906?modelVersionId=28569"> </a><a target="_blank" rel="ugc" href="https://civarchive.com/models/267710">Furception</a> [SD-WebUI\models\VAE]</li></ul>Hires. fix<ul><li>Hires steps = Steps * Denoising strength</li><li>Denoising strength = 0.25</li><li>Hires upscaler = <a target="_blank" rel="ugc" href="https://mega.nz/folder/qZRBmaIY#nIG8KyWFcGNTuMX_XNbJ_g/folder/zJI3EY6T">4x-UltraMix_Smooth </a>[SD-WebUI\models\ESRGAN]</li></ul>ControlNet<ul><li>ControlNet = softedge_hed, <a target="_blank" rel="ugc" href="https://huggingface.co/lllyasviel/ControlNet-v1-1/tree/main">control_v11p_sd15_softedge</a></li><li>ControlNet SDXL = softedge_hed, <a target="_blank" rel="ugc" href="https://huggingface.co/SargeZT/controlnet-sd-xl-1.0-softedge-dexined/tree/main">sdxlsoftedge-dexined, </a><a target="_blank" rel="ugc" href="https://civarchive.com/models/929685/noobai-xl-controlnet">noobai-xl-controlnet</a></li><li>ControlNet Weight = 0.35~0.5</li><li>ControlNet Pixel Perfect = true</li></ul>Wan 2.2 ComfyUI Animeted [<a target="_blank" rel="ugc" href="https://mega.nz/file/hMxhGBoT#p67B8GNYXJXGK9GWAgt39ENR_ec2X3tTKPcFGuZuvAk">workflow</a>]<ul><li>model = <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/blob/main/split_files/diffusion_models/wan2.2_ti2v_5B_fp16.safetensors">wan2.2_ti2v_5B_fp16</a></li><li>clip = <a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/WanVideo_comfy/blob/main/umt5-xxl-enc-bf16.safetensors">umt5-xxl-enc-bf16</a> / <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.1_ComfyUI_repackaged/blob/main/split_files/text_encoders/umt5_xxl_fp8_e4m3fn_scaled.safetensors">umt5_xxl_fp8_e4m3fn_scaled</a></li><li>vae = <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/Wan_2.2_ComfyUI_Repackaged/blob/main/split_files/vae/wan2.2_vae.safetensors">wan2.2_vae</a></li><li>lora (turbo) = <a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/WanVideo_comfy/blob/main/LoRAs/Wan22-Turbo/Wan22_TI2V_5B_Turbo_lora_rank_64_fp16.safetensors">lora_wan2.2_ti2v_5B_turbo_lora_rank_64_fp16</a></li><li>lora (furry-nsfw) = <a target="_blank" rel="ugc" href="https://civarchive.com/models/1827161/furry-nsfw-wan-22-5b">wan2.2-5B_furry-nsfw-v2.0-e83</a></li><li>lora (livewallpaper) = <a target="_blank" rel="ugc" href="https://civarchive.com/models/1827161/furry-nsfw-wan-22-5b">wan2.2-5B_</a><a target="_blank" rel="ugc" href="https://civarchive.com/models/1264662?modelVersionId=2080103">livewallpaper-720p</a></li><li>sampler = steps:4, cfg:1, sampler:"eular_ancestral", schedule:"sgm_uniform"</li></ul>SD WebUI<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/datasets/Xeno443/reForge-portable">reForge</a></li><li><a target="_blank" rel="ugc" href="https://huggingface.co/TikFesku/sd-webui-forge-neo-portable/tree/main">Forge</a></li><li><a target="_blank" rel="ugc" href="https://github.com/AUTOMATIC1111/stable-diffusion-webui">A1111</a></li></ul>LoRA Training SDXL<ul><li>imgs count = 15~50</li><li>total steps = epoch * imgs count * folder loop = 3000~4500</li><li>network_dim = 64</li><li>network_alpha = 128 (SDXL) \ 16 (Noob)</li><li>learning_rate = 0.0002~0.0005</li><li>unet_lr = 0.0001 #learning_rate/2</li><li>text_encoder_lr = 0.00005 #learning_rate/4</li><li>lr_scheduler = "cosine_with_restarts"</li><li>mixed_precision = "bf16"</li><li>optimizer_type = "Adafactor"</li><li>optimizer_args = [ "scale_parameter=False", "relative_step=False", "warmup_init=False", ]</li></ul><h3 id="yiffymix-v4x-v-pred-setting">YiffyMix v4x V-pred Setting</h3># YiffyMix v4x A1111 SD-WebUI or Forge V-pred Setting1.Download config file ".yaml" paste next to the model.2.Raname ".yaml" same as model name. ( check the ".yaml" has [parameterization: "v"] line )3.Restart SD-WebUI ( if config fails to load, the model just generate noise )# YiffyMix v4x ComfyUI or StableSwarmUI V-pred SettingAdd node "ModelSamplingDiscrete", Sampling = "v_prediction"<a target="_blank" rel="ugc" href="https://mega.nz/file/9JhWjD6Q#JWe-XhGubwX3BapQfP8A-1Q0S-YU1D1G1z176VxM3U0">workflow demo</a># ComfyUI V-pred Setting (old version)1.Copy config ".yaml" to [ComfyUI\models\configs] and refresh ComfyUI2."Load Checkpoint (With Config)" [Right Click\AddNode\advanced\model_merging]workflow:[Load Checkpoint (With Config)]-[KSampler]-[VAE Decode]-[Save Image]# v-pred mode troubleshootIf you use new version webui-forge and <a target="_blank" rel="ugc" href="https://github.com/lllyasviel/stable-diffusion-webui-forge/issues/123">failed detect v-pred model</a>.Try update webui-forge (click update.bat).When loading v-pred model, you will see this line in cmd console:left over keys: dict_keys(['v_pred'])# v4.x some time will get little fried issue image in this caseUse "Dynamic Prompts with __wildcards__ prompts" and "batch size &gt; 1"Just generate again with same prompts with parament, it will return normal.<h3 id="version-info">Version Info</h3>v1.x [2D,512~768] old model, unstable, low resolutionv2.x [2D,512~896] e621 dataset + 30~40% SD5 datasetv3.0 [2D,3D,512~1088] large dataset then v2.x version, can do 2D, 3Dv3.1 [2D,3D,Real,512~1088]more realistic, but loss some concept (e621 tag count lower 1000)v3.2 [2D,3D,Real,512~1088]unstable version, use SNR version FluffyRock, more noise detailv3.3 [2D,3D,Real,512~1088]stable version, more detail, more sensitive to promptsv3.4 [2D,3D,Real,512~1088] ※include Fluffy Rock Quality Tags-LoRAstable version, more detail, clear result, reduce some noise (like bush,pattern)v3.5 [2D,3D,Real,512~1088]unstable version, contrast &amp; detailed enhancev3.6 [2D,3D,Real,512~1088]stable version, more sensitive to e621 tagsv3.7 [2D,3D,Real,512~1088]stable version, reduce noise overactivity, fix blurry eyes and finger error, fix failed background depthv4.0 [2D,3D,Real,512~1088]v-pred version, remix from <a target="_blank" rel="ugc" href="https://huggingface.co/zatochu/EasyFluff/tree/main">EasyFluff</a>accuracy anatomy and style with fewer prompt,little dim blue average color with smooth noise, weakly negative prompt issuev4.1 [2D,3D,Real,512~1088]v-pred version, color and low contrast fix, reduced realistic noise.v4.2 [2D,3D,Real,512~1088]v-pred version, more contrast and detail, yelow issue.if you feel too fried use <a target="_blank" rel="ugc" href="https://github.com/Seshelle/CFG_Rescale_webui">Rescale CFG </a>=0.35v4.3 [2D,3D,Real,512~1088]v-pred version, clear artist style, fix most yellow/brown issuethe version doesn't need CFG Rescalev4.4 [2D,3D,Real,512~1088]v-pred version, no more yellow/brown issue.the version doesn't need CFG Rescalev5.0 [2D,3D,Real,896~1536]SDXL-lightning version, base on Compassmix XL.v5.1 [2D,3D,Real,896~1536]SDXL-lightning version, more e621 dataset, lower human face, better NSFW stuff.v5.2 [2D,3D,Real,896~1536]SDXL-lightning version, increase average quality. Little unstable than v51 but more creative.v6.0 [2D,3D,res:896~1536]More character better sex pose, limited and uncontrollable style (mostlly anime style).v6.1 [2D,3D,Real,896~1536]More realistic detail, reduce anime style, fix flat and boring background.v6.1a-RE [2D,3D,Real,896~1536]Same as v61 but adjust average style to simi-realistic.v6.2 [2D,3D,Real,896~1536]more effective prompt, keep style also with good realistic detail, saturation down little bit.v6.3 [2D,3D,Real,896~1536]Detail(noise) level between v62 and v61, upgrade character/artist accuracy.v6.4 [2D,3D,Real,896~1536]Upgrade light quality and average detail.<h3 id="note-about-noobxl-to-create-stable-furry-(v6x):">Note about NoobXL to create stable furry (v6x):</h3>use "furry" tag in propmt can let AI create furry not human.use "no human" tag in propmt can avoid NoobXL model add human and reduce anthro affect (more original style).use "anime style" in negative propmt can reduce classic booru style and more realistic.<h3 id="some-sd-prompt-trick:">Some SD prompt trick:</h3>Combine two character:characterA \(characterB\)Avoid tag bleeding:(chain:0)-link fence(cowboy:0) shothigh (collar:0)Multi-tag combine, enhance and reduce token use:from side + side view = from side viewcrossed legs + legs up = crossed legs up<h3 id="basic-style">Basic Style</h3>Negative Prompt SD1.5unusual anatomy, mutilated, malformed, watermark,amputee, mosaic censorship, sketch, monochromeNegative Prompt SDXLmalformed, worst quality, bad quality, signature, text, url<h3 id="3d-artwork-style">3D Artwork Style</h3>Prompt v5x: blender \(software\), ray tracing, 3d, unreal engine<h3 id="photorealistic-style-sdxl">Photorealistic Style SDXL</h3>Prompt v5x:by Mandy Disher, by Wim Wenders, by Robert Rauschenberg, [by Fossa666::0.65],ultra realistic, photorealism, photographNegative prompt v5x: sketch, manga, vector, line art, toonyPrompt v6x: film photography, photorealistic, film grainNegative prompt v6x: anime style, vibrant, pastel<h3 id="photorealistic-style-sd1.5">Photorealistic Style SD1.5</h3>Prompt sd15: [:photorealistic, analog style, realistic, photorealism:0.5]Negative prompt sd15: [:<a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/84134?type=Training%20Data">bwu</a>:0.5]Recommend Refiner: <a target="_blank" rel="ugc" href="https://civarchive.com/models/34469?modelVersionId=328557">IndigoFurryMix v110-Realistic </a>Switch at 0.5

YiffyMix

Recipes v52:partA = Compassmix XL + Boltning v1.0 * 0.35partB = (partA + BananaStrike XL - SDXL Lightning 4 step) * trainDiff:0.25YiffyMix v52 = (partB + Compassmix XL - SDXL Lightning 4 step) * trainDiff:0.25

Files with SHA256: 293a85966be6e6ebeefa1f226c0c59145d1cbae2de3839547690dbd5ca3406d5