Astolfomix XL

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

AstolfoMix-XL是一個文本到圖像的模型，它整合了多種算法和模型進行融合，旨在生成高質量的圖像。該項目基於Stable Diffusion技術，使用了多種合併算法，為圖像生成提供了更多的可能性。

🚀 快速開始

本項目的預覽由 README.md 的頂部會話提供，並通過 convert_sdxl_to_diffusers.py 將SDXL模型從A1111獨立文件轉換為diffuers。設置可能並非最優（無CFG / PAG / FreeU等）。一旦上傳主模型文件，將立即替換預覽擴散器。

✨ 主要特性

• 即使在 SD1 和 SD2 上有經驗，合併問題仍未解決。
• 自 “Extended-FP64” 以來，使用了專用合併器 sd-mecha，它是 mergekit 的SD版本。

📦 安裝指南

文檔未提及安裝步驟，故跳過此章節。

💻 使用示例

基礎用法

文檔未提及基礎用法的代碼示例，故跳過此章節。

高級用法

文檔未提及高級用法的代碼示例，故跳過此章節。

📚 詳細文檔

255c

• 從215c重新開始，增加了40個模型。實現了偽隨機權重。
• 當前版本：x255c-AstolfoMix-25022801-1458190.safetensors
• 推薦版本："x255c"
• 推薦CFG：6.0 (CFG++, SEG 11.0, PAG = 1.0)
• 該版本不太能響應提示。建議作為基礎模型進行訓練。

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

• 附加腳本，附加合併器

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

• 推薦版本："x215c"

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

• 提示詞最少，甚至可以為空。

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合併

• 配方，端到端合併器。

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

• 推薦版本："x215a"

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

• 提示詞最少，甚至可以為空。

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

• 從TGMD擴展而來的版本（從117擴展到192）

• 配方，端到端合併器。

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

• 推薦版本："x191a"

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

• 提示詞最少，甚至可以為空。

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)

• 對116個SDXL模型進行 TGMD合併，未過濾。TGMD是從 Model Stock 修改而來的算法。

• 配方，端到端合併器。

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

• 推薦版本："x116a"

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

• 提示詞最少，甚至可以為空。

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)

• 對102個SDXL模型進行 TSD合併，未過濾。TSD是從 DARE合併 (ICML 2024) 修改而來的算法。

• 配方，端到端合併器。

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

• 推薦版本："x101a"

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

• 提示詞最少，甚至可以為空。

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

• 對73個SDXL模型進行 TIES-SOUP，未過濾。TIES-SOUP是從 TIES合併 (NeurIPS 2023) 修改而來的算法。

• 配方，端到端合併器。

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

• 推薦版本："x72a"

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

• 提示詞最少，甚至可以為空。

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

• 對52個UNETS + (61+42) 個CLIPS進行 均勻合併（從發現的70個模型中）。

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

• 推薦版本："x51"

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

• 提示詞最少。

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

• 對32個UNETS + (19+26) 個CLIPS進行 均勻合併（從21個模型中）。發現的模型數量：42。它是 Uniform Soup 的衍生版本。

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

• 推薦版本："x17" 以獲得完整體驗，或 "x11c" 以聚焦人物。

• 推薦CFG：4.5

• 提示詞最少。

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 / 模型選擇日誌 / 涉及的模型

• 查看 Github上的完整文章

DGMLA概述

• 該算法的空間複雜度應為線性，即O(N)，時間複雜度為O(NlogN)。然而，逐層合併引入了一些常量，影響了實際體驗。

215c概述

• 查看 這篇文章

📄 許可證

本項目採用 creativeml-openrail-m 許可證。