metadata

language:
  - en
license: creativeml-openrail-m
tags:
  - stable-diffusion
  - stable-diffusion-diffusers
  - text-to-image
  - safetensors
inference: true
thumbnail: >-
  https://huggingface.co/6DammK9/AstolfoMix-XL/resolve/main/240234-2213673007-2016-1152-4.5-192-20240210015331.jpg
widget:
  - text: aesthetic, quality, 1girl, boy, astolfo
example_title: example 1girl boy
library_name: diffusers

AstolfoMix-XL (Baseline)

Unsolved merge, even with experience on SD1 and SD2.
See full article in Github

Baseline

Uniform merge of 32 UNETS + (19+26) CLIPS (from 21 models). Discovered model count: 42.
Current version: x17-AstolfoMix-x13te0x14te1.safetensors
Recommended version: "x17" for full experience, or "x11c" for human focused.
Recommended CFG: 4.5
Prompt is minimal.

parameters

(solo:0), (boy:0), (qipao:0.98), [astolfo_\(fate\)], [[lunar new year]], [[[[kowloon]]]]
Steps: 192, Sampler: Euler, CFG scale: 4.5, Seed: 2213673007, Size: 1344x768, Model hash: 82f53a8fe1, Model: x17-AstolfoMix-x13te0x14te1, VAE hash: 26cc240b77, VAE: sd_xl_base_1.0.vae.safetensors, Denoising strength: 0.7, Clip skip: 2, FreeU Stages: "[{\"backbone_factor\": 1.1, \"skip_factor\": 0.6}, {\"backbone_factor\": 1.2, \"skip_factor\": 0.4}]", FreeU Schedule: "0.0, 1.0, 0.0", FreeU Version: 2, Hires upscale: 1.5, Hires upscaler: Latent, Dynamic thresholding enabled: True, Mimic scale: 1, Separate Feature Channels: False, Scaling Startpoint: MEAN, Variability Measure: AD, Interpolate Phi: 0.3, Threshold percentile: 100, Version: v1.7.0

Receipe

Here is a full list of discovered models:

Table with 40+ rows. Click to open.

mega table coming soon

_x01-deepDarkHentaiMixNSFW_v12.safetensors
_x02-animeAntifreezingSolutionXL_v10.safetensors
_x03-hsxl_base_1.0.f16.safetensors
_x04-OpenDalleV1.1.safetensors
_x05-copaxTimelessxlSDXL1_v8.safetensors
_x06-juggernautXL_v8Rundiffusion.safetensors
_x07-kohakuXLBeta_beta7.safetensors
_x08-animagineXLV3_v30.safetensors
_x09-animeboysxl_v10.safetensors
_x10-dreamshaperXL_alpha2Xl10.safetensors
_x11-SDXLRonghua_v40.safetensors
_x12-bluePencilXL_v310.safetensors
_x13-leosamsHelloworldSDXL_helloworldSDXL32DPO.safetensors
_x14-ponyDiffusionV6XL_v6.safetensors
_x15-animagineXL_v20.safetensors
_x16-wdxl-aesthetic-0.9.safetensors
_x17-leosamsHelloworldSDXLModel_helloworldSDXL10.safetensors
_x18-nekoray-xl-1.5m-fp16mixed_e02.safetensors
_x19-nekoray-xl-1.5m-pdg32_e02.safetensors
_x20-explicitFreedomNSFW_beta.safetensors
_x21-nd-run8-weighted-3.safetensors
_x22-kohakuXL_alpha7.safetensors
_x23-nekorayxl_v06W3.safetensors
_x24-SwimInLatent-alpha.fp16.safetensors
_x25-sd_xl_base_1.0.safetensors
_x26-enjoyXLAdvancedEdition_v7AdvancedVersion.safetensors
_x27-explicitFreedomNSFW_alpha.safetensors
_x28-heartOfAppleXL_v10.safetensors
_x29-kohakuXLGamma_rev1.safetensors
_x30-aamXLAnimeMix_v10.safetensors
_x31-counterfeitxl_v25.safetensors
_x32-enjoyPandoraXL_v10.safetensors
_x33-4Guofeng4XL_v12.safetensors
_x34-animeIllustDiffusion_v061.safetensors
_x35-brightprotonukeBPNNo_bpn13.safetensors
_x36-devlishphotorealism_sdxl15.safetensors
_x37-himawarimix_xlV4.safetensors
_x38-aio_v10.safetensors
_x39-sakumix_v10.safetensors
_x40-218xl_.safetensors
_x41-riotDiffusionXLLeagueOfLegendsSplash_v20.safetensors
_x42-himawarimix_xlV5.safetensors
_x43-SDXLAnimeBulldozer_v20.safetensors
_x44-kohakuXLGamma_rev2.safetensors
_x45-awmo-xl.safetensors
_x46-realvisxlV30Turbo_v30Bakedvae.safetensors

Selected models

te0: Models selected for using the CLIP, or conditioner.embedders.0.
te1: Models selected for using the CLIP-AUX, or conditioner.embedders.1.
te2: Models looks acceptable with both CLIPs applied. Not for selection, just for reference.
=sd: Models has the exact same CLIP pair as the original SDXL _x25.
-- is used for placeholder because I need to figure out the model pairs to be merged.
Total / subtotal count will be shown as =xx.
xxa e.g. 09a is the merge of _x17 and _x18, because of model deviation. It introduce some error on expected model weight, but will be negligible when the scale is large.

te0: --,--,03,05,--,10,11,--,--,16,17,09a,10a,20,11a,22,23,24,25,26,27,32,--,36,--,--,41,--=19
te1: 01,02,03,05,06,10,--,12,14,16,17,09a,10a,20,11a,22,--,24,25,26,27,32,33,36,37,38,41,42=26
te2: 01,--,03,05,06,10,--,--,--,16,17,09a,10a,20,11a,22,--,24,25,--,27,32,33,36,37,38,41,--=21
=sd: --,--,03,--,--,10,--,12,--,16,--,-18,-19,--,---,--,--,--,25,--,--,--,--,--,--,--,--,--=7

Excluding the =sd for the actual merge will be shown below. Intermediate model name will be x01te1-AstolfoMix-_x01_x02.

te0:--,--,05,--,11,--,--,17,20,11a,22,23,24,25,26,27,32,--,36,--,--,41,--=14
te1:01,02,05,06,--,12,14,17,20,11a,22,--,24,25,26,27,32,33,36,37,38,41,42=21

unet: Models selected for using the UNET.
nsfw: Models selected for using the UNET. This time I used a NSFW prompt for test.
To keep it easy to trace (originaly it is 33 / 15), I adjust it to be a perfect 2^n. It will be handy for future merge, when new model appears.
x01a: Model pairs which are already available for merge. Therefore I can reuse for saving disk space, a lot.

unet: 01,02,06,07,09,10,11,12,13,15,16,17,20,21,22,23,25,26,27,28,30,31,32,33,34,--,37,38,39,40,42,43,44=32
nsfw: 01,--,06,07,--,--,11,12,13,15,--,--,20,21,22,23,--,--,27,--,--,31,--,--,--,36,--,--,--,40,--,43,--=16
x01a: 01,02,--,--,09,10,11,12,--,15,16,--,--,21,22,--,25,26,27,28,--,31,32,33,34,--,37,38,39,40,--,43,44

Finally here is the result model name:

base=x17-AstolfoMix-x13te0x14te1.safetensors
nsfw=x11c-AstolfoMix-x13te0x14te1.safetensors