Astolfomix XL

🚀 AstolfoMix-XL (255c / 215c / DGMLA-216 / TGMD-192 / TGMD / TSD / TIES-SOUP / Extended-FP64 / Baseline)

AstolfoMix-XL是一个文本到图像的模型，它整合了多种算法和模型进行融合，旨在生成高质量的图像。该项目基于Stable Diffusion技术，使用了多种合并算法，为图像生成提供了更多的可能性。

🚀 快速开始

本项目的预览由 README.md 的顶部会话提供，并通过 convert_sdxl_to_diffusers.py 将SDXL模型从A1111独立文件转换为diffuers。设置可能并非最优（无CFG / PAG / FreeU等）。一旦上传主模型文件，将立即替换预览扩散器。

✨ 主要特性

• 即使在 SD1 和 SD2 上有经验，合并问题仍未解决。
• 自 “Extended-FP64” 以来，使用了专用合并器 sd-mecha，它是 mergekit 的SD版本。

📦 安装指南

文档未提及安装步骤，故跳过此章节。

💻 使用示例

基础用法

文档未提及基础用法的代码示例，故跳过此章节。

高级用法

文档未提及高级用法的代码示例，故跳过此章节。

📚 详细文档

255c

• 从215c重新开始，增加了40个模型。实现了伪随机权重。
• 当前版本：x255c-AstolfoMix-25022801-1458190.safetensors
• 推荐版本："x255c"
• 推荐CFG：6.0 (CFG++, SEG 11.0, PAG = 1.0)
• 该版本不太能响应提示。建议作为基础模型进行训练。

parameters

(car:0), [[mclaren]], (1boy:0), [astolfo]
Steps: 48, Sampler: DDIM CFG++, Schedule type: Automatic, CFG scale: 6, Seed: 3857800394, Size: 1344x768, Model hash: 516c1564c3, Model: x255c-AstolfoMix-25022801-1458190, VAE hash: 235745af8d, VAE: sdxl-vae-fp16-fix.vae.safetensors, Denoising strength: 0.7, Clip skip: 2, Hires upscale: 1.5, Hires upscaler: Latent, SEG Active: True, SEG Blur Sigma: 11, SEG Start Step: 0, SEG End Step: 2048, PAG Active: True, PAG SANF: True, PAG Scale: 1, PAG Start Step: 0, PAG End Step: 2048, Version: v1.10.1

215c

• 部分Git-Rebasin + DGMLA

• 附加脚本，附加合并器

• 当前版本：x215c-AstolfoMix-24101101-6e545a3.safetensors

• 推荐版本："x215c"

• 推荐CFG：6.0 (CFG++, SEG 11.0, PAG = 1.0)

• 提示词最少，甚至可以为空。

parameters

(car:0), [[AMG F1 W11]], (1boy:0), [astolfo]
Steps: 48, Sampler: DDIM CFG++, Schedule type: Automatic, CFG scale: 6, Seed: 2004174654, Size: 1344x768, Model hash: b31b84ff71, Model: x215c-AstolfoMix-24101101-6e545a3, VAE hash: 235745af8d, VAE: sdxl-vae-fp16-fix.vae.safetensors, Denoising strength: 0.7, Clip skip: 2, Hires upscale: 1.5, Hires upscaler: Latent, SEG Active: True, SEG Blur Sigma: 11, SEG Start Step: 0, SEG End Step: 2048, PAG Active: True, PAG SANF: True, PAG Scale: 1, PAG Start Step: 0, PAG End Step: 2048, Version: v1.10.1

DGMLA-216

• DGMLA合并

• 配方，端到端合并器。

• 当前版本：x215a-AstolfoMix-24101101-6e545a3.safetensors

• 推荐版本："x215a"

• 推荐CFG：4.5 (with CHG = 1.0)，3.0 (with PAG = 1.0)

• 提示词最少，甚至可以为空。

parameters

[halloween], (astolfo:0.98), [[[[cemetery]]]]
Steps: 256, Sampler: Euler, Schedule type: Automatic, CFG scale: 3, Seed: 2123095857, Size: 1344x768, Model hash: bdb9f136b6, Model: x215a-AstolfoMix-24101101-6e545a3, VAE hash: 235745af8d, VAE: sdxl-vae-fp16-fix.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: 2, Hires steps: 64, 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, PAG Active: True, PAG SANF: True, PAG Scale: 1, PAG Start Step: 0, PAG End Step: 150, Version: v1.10.1

TGMD-192

• 从TGMD扩展而来的版本（从117扩展到192）

• 配方，端到端合并器。

• 当前版本：x191a-AstolfoMix-24083001-3360d18.safetensors

• 推荐版本："x191a"

• 推荐CFG：4.5 (with CHG = 1.0)，3.0 (with PAG = 1.0)

• 提示词最少，甚至可以为空。

parameters

(car:0), [[lamborghini]], (1boy:0), [astolfo]
Steps: 256, Sampler: Euler, Schedule type: Automatic, CFG scale: 3, Seed: 2435649982, Size: 1344x768, Model hash: 4c118beaa8, Model: x191a-AstolfoMix-24083001-3360d18, VAE hash: 235745af8d, VAE: sdxl-vae-fp16-fix.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 steps: 64, 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, PAG Active: True, PAG SANF: True, PAG Scale: 1, PAG Start Step: 0, PAG End Step: 2048, Version: v1.10.1

TGMD (TIES-GeometricMedian w/ DROP)

• 对116个SDXL模型进行 TGMD合并，未过滤。TGMD是从 Model Stock 修改而来的算法。

• 配方，端到端合并器。

• 当前版本：x116a-AstolfoMix-24060702-01823a9.safetensors

• 推荐版本："x116a"

• 推荐CFG：4.5 (with CHG = 1.0)，3.0 (with PAG = 1.0)

• 提示词最少，甚至可以为空。

parameters

[[striped thighhighs]], [[midriff]], [[striped shirt]], [[hoodie]], [[braid]], [astolfo], [[[[eiffel tower, france]]]]
Steps: 256, Sampler: Euler, Schedule type: Automatic, CFG scale: 3, Seed: 1526207600, Size: 1344x768, Model hash: bc747cafd1, Model: x116a-AstolfoMix-24060702-01823a9, VAE hash: 235745af8d, VAE: sdxl-vae-fp16-fix.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: 2, Hires steps: 64, 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, PAG Active: True, PAG Scale: 1, Version: v1.9.4

TSD (TIES-SOUP w/ DROP)

• 对102个SDXL模型进行 TSD合并，未过滤。TSD是从 DARE合并 (ICML 2024) 修改而来的算法。

• 配方，端到端合并器。

• 当前版本：x101a-AstolfoMix-24050903-4edc67c.safetensors

• 推荐版本："x101a"

• 推荐CFG：4.5 (with CHG = 1.0)，3.0 (with PAG = 1.0)

• 提示词最少，甚至可以为空。

parameters

(hippogriff:0.98), [braid], [[cape]], [[astolfo]], [[[[greece]]]]
Steps: 256, Sampler: Euler, Schedule type: Automatic, CFG scale: 3, Seed: 3134096594, Size: 1344x768, Model hash: 7668681e22, Model: x101a-AstolfoMix-24050903-4edc67c, VAE hash: 235745af8d, VAE: sdxl-vae-fp16-fix.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: 2, Hires steps: 64, 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, PAG Active: True, PAG Scale: 1, Version: v1.9.3

TIES-SOUP

• 对73个SDXL模型进行 TIES-SOUP，未过滤。TIES-SOUP是从 TIES合并 (NeurIPS 2023) 修改而来的算法。

• 配方，端到端合并器。

• 当前版本：x72a-AstolfoMix-240421-feefbf4.safetensors

• 推荐版本："x72a"

• 推荐CFG：4.5 (with CHG = 1.0)，3.0 (with PAG = 1.0)

• 提示词最少，甚至可以为空。

parameters

(car:0), [[mclaren]], [astolfo]
Steps: 256, Sampler: Euler, Schedule type: Automatic, CFG scale: 3, Seed: 1504757665, Size: 1344x768, Model hash: e276a52700, Model: x72a-AstolfoMix-240421-feefbf4, 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 steps: 64, 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, PAG Active: True, PAG Scale: 1, Version: v1.9.3

Extended-FP64

• 对52个UNETS + (61+42) 个CLIPS进行 均匀合并（从发现的70个模型中）。

• 当前版本：x51-AstolfoMix-x60te0x41te1-e2e-240407-feefbf4.safetensors

• 推荐版本："x51"

• 推荐CFG：4.5 (with CHG = 1.0)，3.0 (with PAG = 1.0)

• 提示词最少。

parameters

(car:0), [[mclaren]], (1boy:0), [astolfo]
Steps: 256, Sampler: Euler, Schedule type: Automatic, CFG scale: 4.5, Seed: 1841382272, Size: 1344x768, Model hash: a52eba463d, Model: x51-AstolfoMix-x60te0x41te1-e2e-240407-feefbf4, VAE hash: 26cc240b77, VAE: sd_xl_base_1.0.vae.safetensors, Denoising strength: 0.7, Clip skip: 2, CHG: "{'RegS': 1, 'RegR': 1, 'MaxI': 50, 'NBasis': 0, 'Reuse': 1, 'Tol': -4, 'IteSS': 1, 'ASpeed': 0.4, 'AStrength': 0.5, 'AADim': 2, 'CMode': 'More ControlNet', 'StartStep': 0, 'StopStep': 1}", 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 steps: 64, 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, Refiner switch by sampling steps: True, Version: v1.9.0

Baseline

• 对32个UNETS + (19+26) 个CLIPS进行 均匀合并（从21个模型中）。发现的模型数量：42。它是 Uniform Soup 的衍生版本。

• 当前版本：x17-AstolfoMix-x13te0x14te1.safetensors

• 推荐版本："x17" 以获得完整体验，或 "x11c" 以聚焦人物。

• 推荐CFG：4.5

• 提示词最少。

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

配方 / 模型选择日志 / 涉及的模型

• 查看 Github上的完整文章

DGMLA概述

• 该算法的空间复杂度应为线性，即O(N)，时间复杂度为O(NlogN)。然而，逐层合并引入了一些常量，影响了实际体验。

215c概述

• 查看 这篇文章

📄 许可证

本项目采用 creativeml-openrail-m 许可证。