Type | |
Stats | 3,782 |
Reviews | (391) |
Published | Jun 15, 2024 |
Base Model | |
Hash | AutoV2 24C3AF41A8 |
Pixel Alchemy Hyper H v2.2
I used several workflow approaches to graft clip sd3 onto pixel alchemy, I noticed that I always got the same thing whatever the configuration. T5 doesn't bring any change, most likely this only affects clip G and L.
workflow
https://drive.google.com/file/d/1V_f4icdZfcYYSrqYldEarYA3G4YdR7-W/view?usp=sharing
Hands problem
Hands are currently still a problem, here is a link to my publication on Facebbok, presenting an approach on comfyUI to improve hand renderings based on the Ultralytics Hand Yolov 8s & 8n models :
https://www.facebook.com/groups/stablediffusion/permalink/1399715494027255
Pixel Alchemy Hyper V2.0
Improvement of human body and hands from the merger of the unreleased Pixel Alchemy V2.0 and Hyper-SDXL-12 steps CFG LoRA
Settings (ComfyUI) : Sampler : heunpp2, ddim, Euler/ Scheduler : sgm_uniform, ddim_uniform, simple / Steps : 16 to 20 / cfg : 2.5 to 3
For A1111, I won't be able to give any recommendations, because I don't use it, but I think that with a few tests you will end up finding what works best.
Pixel Alchemy Hyper V1.6
It's a merging of Pixel Alchemy V1.6 model with 2 LoRAs :
Hyper-SDXL-12 steps CFG LoRA : https://huggingface.co/ByteDance/Hyper-SD/blob/main/Hyper-SDXL-12steps-CFG-lora.safetensors
Hand Detail XL LoRA : https://civitai.com/models/260852/hand-detail-xl-lora
Settings (ComfyUI):
Sampler : ddim / Scheduler : ddim_uniform / Steps : 20 / cfg : 3
Sampler : Euler / Scheduler : Simple / Steps : 14 / cfg : 2.5
Sampler : heunpp2/ Scheduler : sgm_uniform / Steps : 18 / cfg : 2.5
but feel free to test others settings, specially for A1111
Pixel Alchemy V1.6
Version 1.6 of Pixel Alchemy brings a significant improvement in the rendering of photorealistic portraits, the texture of the skin appears more natural.
It's a merging of Pixel Alchemy V1.3 with 1 checkpoint and 2 LoRAs
RealVisXl model : https://civitai.com/models/139562?modelVersionId=344487
Steve McCurry - Photography SDXL LoRa : https://civitai.com/models/213283/steve-mccurry-photography-sdxl-lora?modelVersionId=240247
Essenz - Better Photography (Style LoRa for SDXL 1.0) : https://civitai.com/models/198378?modelVersionId=250573
Pixel Alchemy V1.3
It's a merge of 4 checkpoints and 3 LORAs (see list below).
One of the main goals is was to obtain a very versatile checkpoint. It's NSFW one.
Quite efficient in rendering landscapes, realistic portraits, surrealism and gives better results with noise reinjection in ComfyUI (see below, the recommended extension)
All my uploaded images were generated with low strength ADetailer (Eyes / hands / Face).
Models merged in:
Realism Engine SDXL : https://civitai.com/models/152525/realism-engine-sdxl
Crystal Clear XL : https://civitai.com/models/122822/crystal-clear-xl
Juggernaut : https://civitai.com/models/133005?modelVersionId=348913
LoRAs merged in:
Detail Tweaker XL : https://civitai.com/models/122359/detail-tweaker-xl
xl more art-full / xl_real / Enhancer : https://civitai.com/models/124347/xlmoreart-full-xlreal-enhancer
SDXL Film Photography Style : https://civitai.com/models/158945/sdxl-film-photography-style?modelVersionId=180569
ComfyUI workflow used to merge :
https://drive.google.com/file/d/1vcVgmKKS-T1YSKEmHDdB2j_5yGF89T8y/view?usp=drive_link
Presets merger used in workflow :
https://github.com/WASasquatch/ComfyUI_Preset_Merger
Recommended settings :
· Sampler: dpmpp_2m / Euler / heunpp2 / dpmpp_3m_sde_gpu
· Scheduler : Karras / Exponentiel / sgm_uniform
· CFG: 5 - 7
· Steps: 36 - 46
Comfyui workflow example used with Pixel Alchemy :
Noise injection with YANC extension gives better results (https://github.com/ALatentPlace/ComfyUI_yanc)
https://drive.google.com/file/d/1azdll0HG8qHC6lcsArRdITs7VWblKjNi/view?usp=sharing
NOTES :
All photos are generated with ComfyUI and published with their metadata.
option 1 : To display them, you must click on the "i" icon at the bottom right of the image. If this icon does not appear, it is because Civitai cannot read the ComfyUI file correctly.
option 2 : click on the image to open up the image page (not the post page!) then click on the download icon on the "icon/menu bar" above the image to download the original PNG. You can then drop the PNG into your ComfyUI to recover the workflow
Do not hesitate to ask for the metadata in the comments if they do not appear, I will send them to you in the comments