전우애 실시!
This model is a merger of the Comradeship XL v14VZ model and the Experimental NoobAI with Rectified Flow + EQ-VAE model. (https://civitai.com/models/2071356?modelVersionId=2343811)
This model requires Rectified Flow. YOU NEED TO USE COMFYUI/SWARMUI or any webui that can do ModelSamplingSD3 at 2.5
Comradeship XL v14VW3 × WAI-NSFW-illustrious-SDXL v15 Karcher-merge V-pred model
This is a minor update model of comradeshipXL-v14KC.
A V-Pred merge model with slightly improved natural language prompt performance using Karcher-merge. (https://github.com/win10ogod/Karcher-merge)
The base SDXL model is a 1:1 mix of ILXL 2.0 and NoobAI EPS 1.1. This is a specialized model for high resolutions of 1.5K~1.75K.
As a model focused on improving the performance of high-resolution generation in NoobAI-based models, the supported resolutions for this model are not landscape, with a recommended 1536x1536 (1.5K) / up to 1796x1796 (1.75K).
Improved high-resolution generation performance based on v14K1A model.
A variant of Illustrious XL 1.1 from Comradeship XL v14.
comradeshipXL-v17T3 + ((comradeshipXL-v14K3 - Illustrious XL 0.1) * 0.5) = comradeshipXL-v17T4
Merged noobai-XL EPS 1.1 while retaining the 1536x1536 high-resolution capabilities of Illustrious-XL 1.0.
Animagine XL 4.0 Opt + mann-e-004 LoRA = comradeshipXL-v17T3
Animagine XL 4.0 + mann-e-004 LoRA = comradeshipXL-v17T1
Illustrious XL 1.0 version of Comradeship XL v14
This is a minor update model of Comradeship XL v14T15V. Changed NoobAI XL V-Pred 0.75s to NoobAI XL V-Pred 1.0 in merge recipe for Comradeship XL v14T15V model. Since it is not a general SDXL (eps-prediction) model, it must be used in a v-prediction support program.
SDXL Hyper version of Comradeship XL v14V. Available in 6~15 steps / CFG 1.0 settings. Using the recommended settings, you can create images quickly.
This is a minor update model of Comradeship XL v14T14V. In v14T14V, I changed the base model to NoobAI XL V-Pred 0.75s and merged the differences from CLIP-SAE-ViT-L-14 - CLIP-ViT-L-14. Since it is not a general SDXL (eps-prediction) model, it must be used in a v-prediction support program.
This is a v-prediction model based on NoobAI XL V-Pred 0.5 / 0.65s
Since it is not a general SDXL (eps-prediction) model, it must be used in a v-prediction support program.
This model recommends using Euler CFG++ and CFG is recommended between 1 and 2.5. A minimum of 8 steps is required, and 12 to 14 steps are recommended.
In the merge recipe for the Comradeship XL v14T13 model, I changed from the NoobAI-XL Early Access version to the NoobAI-XL 0.5/0.75 version.
This model is based on Illustrious-XL v0.1.
Merge recipes can be found at this link.
https://huggingface.co/hanzogak/comradeshipXL/tree/main/CSv14T1S-KIT
Anime performance has been slightly improved.
Check the ComfyUI EXIF of the models for details on how to merge.
v9B2-Lite6 → v9M6K → v9MF4 → v9MB4
https://huggingface.co/hanzogak/comradeshipXL/tree/main
Hard ???? performance has been improved.
Check the ComfyUI EXIF of the models for details on how to merge.
https://huggingface.co/hanzogak/comradeshipXL/tree/main
v9B2-Lite3 → v9M3 → v9M3K → v9MF → v9MB
The v9M model is a hybrid Pony model that has the character/artist features of the anime SDXL model. Features may not be as clear compared to purebred models.
InterDiffusion-4.0 - SDXL 1.0 = InterDiffusion-4.0 LoRA
Kohaku-XL Epsilon rev1 - kohaku-xl-epsilon-lyco = dummy2
Kohaku-XL Epsilon rev3 - dummy2 = Epsilon_dummy2 LoRA
comradeshipXL-v9AA-SPO-DPO-Flash-MaPObeta + mapo-safe LoRA + mann-e-004 LoRA + (InterDiffusion-4.0 × -1) + (Epsilon_dummy2 LoRA × -1) + marine-animeXL-290001 LoRA = comradeshipXL-v9AD2-Lite
Kohaku-XL Epsilon rev3 + mann-e-004 LoRA + SPO LoCon + DPO LoRA + mapo-safe LoRA + mapo-beta LoRA = comradeshipXL-v10T1
comradeshipXL-v10T1 + marine-animeXL-290001 LoRA = comradeshipXL-v10T3
Neta Art XL v2 + mann-e-004 LoRA + SPO LoCon + DPO LoRA + mapo-safe LoRA + mapo-beta LoRA = comradeshipXL-v11T1
comradeshipXL-v10T3 × 0.33 + comradeshipXL-v11T1 × 0.33 + comradeshipXL-v9AD2-Lite × 0.34 = comradeshipXL-v9B2-Lite
comradeshipXL-v9B2-Lite - SDXL 1.0 = CSv9B2-Lite LoRA
CSv9B2-Lite × 0.4 Unet + AAAAutism AAAA Restart × 0.6 Unet + CSv9B2-Lite TE + CSv9B2-Lite LoRA × 0.35 = comradeshipXL-v9M
CFG 4~6 / euler ascestral / beta
This model is a merge model based on wdv-test-v0.1 from wdv-tests.
ModelSamplingWaifuDiffusionV is required. (ComfyUI ONLY!)
wdv-test-v0.1 - Animagine 3.1 = wdv01 LoRA (Unet only)
Mann-E_Dreams-0.0.4 - SDXL 1.0 = mann-e-004 LoRA (Unet only)
wdv-test-v0.1 + mann-e-004 LoRA × 0.75 + wdv01 LoRA × (-0.01) = comradeshipXL-v12T1-WDV01
ArtiWaifu Diffusion - SDXL 1.0 = ArtiWaifu LoRA
wdv-test-v0.1 + mann-e-004 LoRA × 0.75 + ArtiWaifu LoRA × 0.6 = comradeshipXL-v12T3-WDV01
In this version, mapo-beta LoRA has been additionally merged in v9A.
This version merges SPO, DPO, and Flash's LoRA.
kivotos-xl-2.0 unet × 0.5 + autismmixSDXL_autismmixConfetti unet × 0.5 + kivotos-xl-2.0 Text Encoder = comradeshipXL-v7T10
comradeshipXL-v9T4 × 0.5 + comradeshipXL-v7T10 × 0.5 = comradeshipXL-v9T13
comradeshipXL-v9T13 + SPO LoCon+ DPO LoRA + Flash LoRA × 0.55 = comradeshipXL-v9A-SPO-DPO-Flash
It's a mix of ArtiWaifu, Animagine XL V3.1, and Pony. Haven't tested it, but I don't think LoRA will work.
comradeshipXL-v9T4 × 0.5 + comradeshipXL-v7A × 0.5 = comradeshipXL-v9
Comradeship XL Version 9 Test 4
Hassaku XL (??????) v 1.1 - Pony Diffusion V6 XL = marine-animeXL_h11pd6_locon
PD for Anime - Pony Diffusion V6 XL = marine-animeXL_pa2pd6_locon
autismmixSDXL autismmixConfett + marine-animeXL_pa2pd6_locon + marine-animeXL_h11pd6_locon = comradeshipXL-v8T1
ArtiWaifu Diffusion v1.0 unet × 0.7 + comradeshipXL-v8T1 unet × 0.3 + ArtiWaifu Diffusion v1.0 Text Encoder = comradeshipXL-v9T4
Comradeship XL Version 9 Test 1
ArtiWaifu Diffusion v1.0 unet × 0.7 + autismmixSDXL autismmixConfetti unet × 0.3 + ArtiWaifu Diffusion v1.0 Text Encoder = comradeshipXL-v9T1
This model is intended to combine the performance of pony_pencil-XL with support for Animagine XL V3.1 characters.
U-Net data from PD6 (Pony)-based models are merged. Therefore, some SDXL LoRAs may not work.
comradeshipXL-v7P unet × 0.5 + pony_pencil-XL v1.0.1 unet × 0.5 + comradeshipXL-v7P Text Encoder = comradeshipXL-v7PP
Pony Diffusion V6 XL - sd_xl_base_1.0 = marine-transXL-xl_to_pd6_locon
Animagine XL V3.1 - sd_xl_base_1.0 = marine-transXL-xl_to_am31_locon
comradeshipXL-v7P.json(Animagine XL V3.1, autismmixSDXL_autismmixConfetti, marine-transXL-xl_to_am31_locon, marine-transXL-xl_to_pd6_locon) = comradeshipXL-v7P
comradeshipXL-v7P unet × 0.7 + autismmixSDXL_autismmixConfetti unet × 0.3 + comradeshipXL-v7P Text Encoder = comradeshipXL-v7PT1
comradeshipXL-v7T5 unet × 0.55 + comradeshipXL-v7PT1 unet × 0.45 + comradeshipXL-v7T5 Text Encoder = comradeshipXL-v7PT2
comradeshipXL-v7PT2 + Kohaku-XL Epsilon Rev1 × comradeshipXL-v7A.json = comradeshipXL-v7A
Animagine XL V3.1 + Kohaku-XL Epsilon Rev1 * comradeshipXL-v7T5.json = comradeshipXL-v7T5
comradeshipXL-v7T5 + marine-animeXL_v06v02m * 0.5 = comradeshipXL-v7
Animagine XL V3.1 - Animagine XL V3.0 = marine-animeXL-3130
anima_pencil-XL v2.0.0 + marine-animeXL-3130 + marine-animeXL_v06v02m + (marine-animeXL-d1db × 0.6) = comradeshipXL-v6
Hassaku XL beta v0.6 - Hassaku XL beta 0.2M = marine-animeXL_v06v02m
Kohaku-XL Delta rev1 - Kohaku-XL Delta base = marine-animeXL-d1db
anima_pencil-XL v2.0.0 + (marine-animeXL-d1db × 0.65) + marine-animeXL_v06v02m = comradeshipXL-v6
Pony Diffusion V6 XL + dpo-sdxl-text2image-v1-LoRA + plastic-novel LoRA × 0.25 + plastic-1 LoRA × 0.25 + plastic-2 LoRA × 0.25 + plastic-3 LoRA × 0.25 = comradeshipXL-v5A-DPO
bulldozer_BETA - Animagine XL V3 = marine-animeXL-am3_to_ab1 LoRA
Animagine XL V3 Base + dpo-sdxl-text2image-v1-LoRA + marine-animeXL-310001 LoRA + marine-animeXL-b7b5 LoRA + marine-animeXL-b71b5 LoRA + marine-animeXL-200121 LoRA + marine-animeXL-am3_to_ab1 LoRA = comradeshipXL-v4BB-DPO
Simple LoRA merge model based on Pony Diffusion V6 XL.
Pony Diffusion V6 XL + dpo-sdxl-text2image-v1-LoRA + Smooth Anime Styles for Pony Diffusion V6 XL LoRA + Anime Styles for Pony Diffusion V6 XL LoRA = comradeshipXL-v5Lite-DPO
Changed Animagine XL V3 to Animagine XL V3 Base
blue_pencil-XL v3.1.0 - blue_pencil-XL v0.0.1 = marine-animeXL-310001 LoRA
Animagine XL V3 + dpo-sdxl-text2image-v1-LoRA + marine-animeXL-310001 LoRA + marine-animeXL-b7b5 LoRA + marine-animeXL-b71b5 LoRA + marine-animeXL-200121 LoRA = comradeshipXL-v4-DPO
Changed blue_pencil-XL v2.9.0 to v3.1.0
AnySomniumXL v2 - AnySomniumXL 1.2.1 = marine-animeXL-200121 LoRA
DucHaiten-Real3D-NSFW-XL v1.0 - v0.16 = marine-beautyXL-100016 LoRA
Hassaku XL beta v 0.1 + dpo-sdxl-text2image-v1-LoRA + marine-animeXL-290001 LoRA + marine-beautyXL-100016 LoRA + marine-animeXL-b7b5 LoRA + marine-animeXL-b71b5 LoRA + marine-animeXL-200121 LoRA = comradeshipXL-v3C-DPO
Kohaku-XL beta7.1(?) - Kohaku-XL beta5 = marine-animeXL-b71b5 LoRA
blue_pencil-XL v2.9.0 - blue_pencil-XL v0.0.1 = marine-animeXL-290001 LoRA
Hassaku XL beta v 0.1 + dpo-sdxl-text2image-v1-LoRA + marine-animeXL-290001 LoRA + marine-animeXL-b7b5 LoRA + marine-animeXL-b71b5 LoRA = comradeshipXL-v3-DPO
Baguette-Anime XL v1.0 - SDXL 1.0 = marine-transXL-xl_to_ba
comradeshipXL-v3-DPO + (marine-transXL-xl_to_ba * 0.6) = comradeshipXL-v3B-DPO
Changed blue_pencil-XL v2.0.0 to v2.9.0
Kohaku-XL beta7 - Kohaku-XL beta5 = marine-animeXL-b7b5 LoRA
blue_pencil-XL v2.0.0 - blue_pencil-XL v0.0.1 = marine-animeXL-200001 LoRA
Hassaku XL beta v 0.1 + dpo-sdxl-text2image-v1-LoRA + marine-animeXL-200001 LoRA + marine-animeXL-b7b5 LoRA = comradeshipXL-v2-DPO
The license of comradeshipXL-v4/v6/v7 is the same as Animagine XL V3.
The license of comradeshipXL-v5 is the same as Pony Diffusion V6 XL.
This model is a merger of the Comradeship XL v14VZ model and the Experimental NoobAI with Rectified Flow + EQ-VAE model. This model requires Rectified Flow. YOU NEED TO USE COMFYUI/SWARMUI or any webui that can do ModelSamplingSD3 at 2.5
1. Modèle partagé uniquement à l'apprentissage et au partage. Droits d'auteur et interprétation finale réservés à l'auteur original.
2. Auteur souhaitant revendiquer le modèle : Contactez officiellement SeaArt AI pour l'authentification. Nous protégeons les droits de chaque auteur. Cliquer pour revendiquer
