ND911

I do not know what the difference is between Annealed, but I do prefer it between the two. GGUFS all created with latest llama.cpp, gguf, convert and patches as of today

EclecticEuphoriaProjectChimera 50 Chroma finally finished up it's training and here we are 50! So we have a 48, 50, Krea, Flux mixed in several ways and methods to blend them all together to be pretty universal in diversity, does NSFW, and use plain language or Danboruu tags work. Images should contain all metadata for workflows in comfyui. I did not create quants lower than Q4 as the quality really dropped off. Model has been updated with Chroma 48 and some small tweaks, No Flux.1-dev was used in this one. EclecticEuphoriaProjectChimera - Initial Release This is a mix model of shuttleai/shuttle-jaguar, lodestones/Chroma, and an unpublished Flux model. I wasn't expecting much, but was surprised by the results right out of the gate, so tossing it up here for funs. Doesn't do flux lora's tho, maybe chroma loras. I don't use loras much, not even sure if chroma loras exist. Can do NSFW, just not sure to what extreme as I don't do much NSFW stuff. Examples and WorkFlows available and all images should have workflow embedded

A Flux model that doesn't look like a flux model in Guidance mode. Never a fan of the flux-dev look. All images in examples have simple workflow I uploaded a v4 last night and I noticed it was causing too much banding issues with varioius samplers so I pulled it and reforged it a bit to eliminate the banding. All images should have workflow embedded. Styles - more added, bolder, better adhearance, better clip fp32, and a WF that produces. I noticed a strong nipple bleed through clothing with V1 while doing illustration images. I did a block by block model merge to try and correct the problem and seeing good results

A experimental mix of EclecticEuphoriaProjectChimera50 and Flux1-Dev-SRPO-v1-fp16. Seems to be pretty universal. I mainly used the fp8 to create images. Each image should have the workflow and workflows uploaded also. In testing several models I have found seeds3 with beta57 produce great prompt adhearance and works best for words.

GGUFs for FLUX Monte Carlo (Full FP32) Go Follow and Give Felldude some love for his hard work.

The workflow combines all controlnet models into one and controlled by switches. Also include Ksampler and Custom Sampler with Demon Details, simply use the switch to choose your sampler. The SD 3.5 model I used in the workflow is the SD 3.5 SLGWeighted Merge Example Images of SD 3.5 SLGWeighted Merge on Civitai Clip models for GGUF uploaded and go into the ComfyUI\models\clip Other Clip models uploaded that are used in the workflow also Controlnet models go into the ComfyUI\models\controlnet

A Flux model that doesn't look like a flux model in Guidance mode. It also works without Guidance, so either way works. Never a fan of the flux-dev look. All images straight from Sampler and have workflow embedded. Model can have nipple bleed thru on to clothes

--- tags: - art license: apache-2.0 SD1.5forrealsv11 v11 builds on to v10 adding film grain and details to get more realistic. I use the model as a refiner for other models to bring the realistic to the image. v10 is a bit of an experiment merging it with goddessProjectLite. WF in the image. I use the model as a refiner for other models to bring the realistic to the image. SD1.5forrealsv8 Uploaded more examples and a WF for v8, images in examples should have workflows embeded for comfyui. I use the model as a refiner for other models to bring the realistic to the image. SD1.5forrealsv4 Straight to v4. This model series is my favorite refine/companion for Flux, but with k2, I wanted just a bit more from the model. k3 was born, but it was a bit too strong for my purpose so v4 is a better blend of the k2 n k3. - adds just a bit more of details - adds just a bit more of contrast "A portrait of a woman captured in a side profile. She has long, wavy, dark hair that falls to her shoulders. Her skin is smooth, and her eyes are a striking shade of blue. The background is a muted gray, contrasting with her skin and hair. The woman's pose is relaxed; one hand rests on her head, and the other rests gently on her shoulder. The image is realistic, emphasizing the woman's natural beauty and the (texture of her skin). Hyper-realistic, Skin Texture, Photorealistic, National Geographic, Time Magazine, Life Magazine, Masterpiece, Award-winning photography, Ultra-detailed, Microscopic, High-resolution, 4K, 8K, 16K, Realistic, Authentic, Accurate, Highly detailed, Extremely detailed, Incredibly detailed, Real life" I kinda made this one by accident, wanting a realistic looking refiner, turns out this does the job well

EclecticEuphoriaIllusRealv3 A tweak or two under the hood of v2 - workflow in the images EclecticEuphoriaIllusRealv2 Continuing on with the push for realistic, spent a lot of time working with a workflow to compliment a realistic image. I went with the ClownShark Sampler this time. The model works with all sorts of various samplers and schedulers A Heavy merge of realistic models from illustrious, pony, sdxl. Workflow in the images, also uploaded workflows. Workflow is moduluar, just turn off and turn on what you want to use. It will do NSFW

I thought I had uploaded this model in the past, but apparently not. This is a simple Illustrious merge to come up with this styled look. Images should contain workflow for comfyui.

This is the SD1.5forreals with anime slapped into it. See the examples folder