You can get KreaLens1, Gwen3, KreaLife1, KreaEdit1, Klein4, Cat1 or Hyper1 for $10 and Ultra5, Zuna1 or Zia1 for $5.50 on Ko-Fi. 30-day membership subscriptions available there too. Any problem? Need another payment option? Contact me via chat.
To create up to 85 free images per day in highest quality you can use Big Love here. For discussing photorealistic image generation join me on Discord.
KreaEdit/KreaLife/KreaLens Big Love are a finetunes of Krea 2 that can do txt2img and NSFW image editing. It is more photorealistic, a lot more NSFW-capable and has better skin texture than the original Krea 2. They produce only photorealistic images with booru tags (not anime like other Krea 2 models). They offers the best anatomy (with the least problems) of all Big Love checkpoints. KreaEdit/KreaLens produces more professional style images while KreaLife is trained more on amateur & candid images.
Move the model file into the diffusion_model sub folder of the models folder of ComfyUI, qwen3vl_4b_bf16.safetensors (12+ GB VRAM) or qwen3vl_4b_fp8_scaled.safetensors (6-8 GB VRAM) into the text_encoder sub folder and qwen_image_vae.safetensors into the vae sub folder. To make the GGUF version work in ComfyUI, you need to unzip the second .zip file from Optional Files into the custom_nodes sub folder.
For editing they require the Krea2Edit nodes that can be installed via Comfy Manager by searching for "Krea 2 Identity Edit". For more face consistency try increasing the grounding_px parameter to 1024 and ref_boost to 3 or 4.
Settings are 8 steps (3-4 steps for upscaling), cfg1, er_sde sampler, Beta scheduler, 1024x1536 pixels. Drag'n'drop showcase images into ComfyUI for a workflow.
Big Love Gwen is a finetune of Qwen Edit 2511 and can do NSFW image editing. It produces more realistic and sharp images with better skin texture & genitals than other Qwen checkpoints/loras. Txt2img is lower quality, but generates nice results too.
It is recommended to use ComfyUI. The Gwen workflows offer a skin enhance feature that is not available in Forge Neo. Move the Gwen model file into the diffusion_model sub folder of the models folder of ComfyUI, qwen_2.5_vl_7b_fp8_scaled.safetensors into the text_encoder sub folder and qwen_image_vae.safetensors into the vae sub folder.
For training a lora read here.
Settings are 6-8 steps (3 steps for upscaling), cfg 1-1.5, Euler sampler, Beta scheduler, 832x1216 or 1024x1536 pixels. Drag'n'drop showcase images into ComfyUI for a workflow.
Big Love Klein is a fine-tune of Flux 2 klein. It can do txt2img and also edit images with prompts. It only needs 4 steps, so is quite fast. Check the "About this version" box on the right for info on the individual variants. Some are up to 3x faster.
It is recommended to use it with ComfyUI. Put the downloaded model into the diffusion models sub folder of ComfyUI. You also need qwen_3_8b_fp8mixed.safetensors in the text_encoders sub folder (Qwen3-8B-Q5_K_M.gguf wth 8 GB VRAM) and flux2-vae.safetensors in the vae sub folder. Check out these anatomy tips.
Settings are 4 steps (2 steps for upscaling), cfg1, Euler sampler, Beta or Simple scheduler, 832x1216, 1024x1536, 1280x1920 or 1536x2240 pixels. Drag'n'drop showcase images into ComfyUI for a workflow.
Big Love Cat is a finetune of LongCat and allows NSFW image editing. It reproduces faces extremely well, with almost every image, so a lot more reliable than Klein & Gwen. It is only 6B and needs 8 steps cfg 1. Same Apache 2.0 license and text encoder as Qwen Image but 8 GB VRAM are enough. Txt2img and sex scenes are still experimental. The base model was only trained on one input image & only picks up random details from a 2nd image. The bf16 version of Cat is faster than Klein & Gwen. The mxfp8 version is 1.6x faster than bf16 and just as fast as Gwen.
There is a new detailer workflow below that I developed for Cat but can be used with any model's images. It uses Photo, Ultra, Hyper or Klein for detailing nipples and pussies.
It is recommended to use ComfyUI. Move the model file into the diffusion_model sub folder of the models folder of ComfyUI, qwen_2.5_vl_7b_fp8_scaled.safetensors into the text_encoder sub folder and ae.safetensors into the VAE sub folder.
Training speed is more than 2x faster than any transformer model that I know. For training a lora you need SimpleTuner, which requires WSL under Windows and has a WebUI similar to AI-Toolkit. Detailed instructions will become available soon...
Big Love Cat produces the sharpest images with er_sde/Normal. In a few cases this adds subtle color grain which can be avoided by increasing steps to 12. er_sde/beta and Euler produce softer images without grain.
Settings are 8-12 steps (4 steps for upscaling), cfg 1, er_sde/Normal sampler/scheduler (Softer: Euler/Beta or er_sde/Beta), 832x1216, 1024x1536 or 1232x1824 pixels. Drag'n'drop showcase images into ComfyUI for a workflow.
Big Love Eryn is a fine-tune of Ernie-Image(-Turbo). You can run it locally with ComfyUI and Forge Neo. The fp8 version works with 8 & 12 GB VRAM. Use the (pruned) bf16 version if you have 16 GB VRAM or more. The "Full" versions are only meant for training a lora (coming soon!). Put the downloaded model into the diffusion models sub folder of ComfyUI or the StableDiffusion sub folder of Forge Neo. You additionally need to place ministral-3-3b.safetensors in the text_encoder sub folder and flux2-vae.safetensors in the VAE sub folder.
Settings are 4-6 steps (3-4 steps for upscaling), cfg1, Euler sampler, Beta scheduler, 832x1216 or 1024x1536 pixels. Drag'n'drop showcase images into ComfyUI for a workflow.
Big Love Hyper supports generating at image sizes of 1664x2432 (4 megapixel) and higher with SDXL. Can be easily upscaled to 9 megapixel unlike the normal low-res images. Hyper currently has not fully stable anatomy, so needs more runs. Some prompts work better than others. It produces highly detailed and life-like images close to real photos. It has a special quality that cannot be achieved by just upscaling. The Hyper versions of Big Love are pay-only and special license conditions apply (see below).
Settings are DMD2 lora with strength 1.0, LCM Exponential, 8 steps, cfg 1, clip skip 2, 1664x2432 or 1536x2240. Upscale with img2img with DMD2 lora with strength 1.0, same prompt, LCM Exponential, 1.25-1.5x, 4-8 steps, cfg 1, 0.3-0.5 denoise, clip skip 2. Look below for a ComfyUI workflow.
Big Love ZT, Zia or Zuna are fine-tunes of Z-Image(-Turbo) on the Big Love dataset. You can run it locally with ComfyUI and Forge Neo. The fp8 version works faster with 8 GB VRAM but produces more noise so may sometimes look more realistic, but details are not as good. Use the (pruned) bf16 version if you have 12 GB VRAM or more. The "Full" version is only meant for training a lora. Put the downloaded model into the diffusion models sub folder of ComfyUI or the StableDiffusion sub folder of Forge Neo. To make Big Love ZT/Zia/Zuna work you additionally need to place qwen3_4b.safetensors in the text_encoder sub folder and ae.saftendors in the VAE sub folder.
Settings are 8 steps, cfg1, Euler or DPM++ 2s a RF sampler, Normal or Simple or Beta scheduler, 832x1216, 1024x1536 or 1280x1920 pixels. For upscaling 8 steps, cfg1, Euler sampler, Normal or Simple or Beta scheduler, 1.5x upscale. Here is a ComfyUI workflow embedded in a image.
Big Love Ultra supports generating at image sizes of 1280x1870, 1360x1984 and higher. Can be easily upscaled to 6 megapixel unlike the normal low-res images. More photorealistic and detailed. A special quality that cannot be achieved by just upscaling. Ultra produces more details & sharpness when upscaling and detailing low-res images. The Ultra versions of Big Love are pay-only and special license conditions apply. See below.
Settings are DMD2 lora with strength 1.0, LCM Exponential, 8 steps, cfg 1, clip skip 2, 1280x1864 or 1360x1984. Upscale with img2img with DMD2 lora with strength 1.0, same prompt, LCM Exponential, 1.25-1.5x, 4-8 steps, cfg 1, lora strength 1.0, 0.3-0.5 denoise, clip skip 2. Look below for a ComfyUI workflow.
Big Love Photo and Insta1 are finetuned versions of XL. Lust1 is a finetuned version of Lustify merged with Photo. They were trained on thousands of images and dozens of new concepts. Why they are the most realistic and versatile. For more details and prompting Photo1 read this article.
Settings are DMD2 lora with strength 1.0, LCM Exponential, 8 steps, cfg 1, clip skip 2, 832x1216 or 1024x1496. Upscale with img2img with DMD2 lora, same prompt, LCM Exponential, 1.25-1.5, 4-8 steps, cfg 1, lora strength 1.0, 0.3-0.5 denoise, clip skip 2. Look below for a ComfyUI workflow.
Big Love XL is a combination of 5 different training branches of SDXL: original SDXL, bigASP, NatVis, Anteros and Pony. With a bit of Pony in it, some Pony loras (poses, characters) work too. It outputs a more photorealistic & creative look than the Pony versions and can also do paintings & cartoon.
Settings are DMD2 lora with strength 1.0, LCM Exponential, 8 steps, cfg 1-1.5, clip skip 2, 832x1216 or 1024x1496. Upscale with img2img with DMD2 lora, same prompt, LCM Exponential, 1.5-2x, 8-12 steps (XL1/XL2), 4-8 steps (XL2.5-XL4), cfg 1, lora strength 1.0, 0.3-0.5 denoise, clip skip 2. Look below for a ComfyUI workflow.
Big Love Pony produces different images than both Pony & SDXL-based models. Pony tags are understood, but interpreted a bit differently. Some images lean more in the one or other direction depending on the prompt. It works with Pony loras as well as normal SDXL ones. It can turn anime/cartoon into photorealistic images in img2img. Pony3 includes a bit of Illustrious.
Settings are DMD2 lora with strength 1.0, LCM Exponential, 8 steps, cfg 1-1.5, clip skip 2, 832x1216 or 1024x1496. Upscale with img2img with DMD2 lora, same prompt, LCM Exponential, 1.5-2x, 8-12 steps, cfg 1, lora strength 1.0, 0.3-0.5 denoise, clip skip 2.
Generation Workflows
Download one of the following images and drag'n'drop it onto ComfyUI:
Z-Turbo/Z-Base txt2img & Upscale Workflow (recommended)
SDLX DMD2 txt2img & Upscale Workflow (recommended)
SDXL DMD2 txt2ing & Upscale for Big Love Ultra
SDXL DMD2 img2img Workflow
SDXL Lightning Workflow
Workflows for all other models are embedded in the showcase images above.
Detailer Workflows
Here are the detailer workflows that I use myself. Drag'n'drop the images linked below into Comfy:
Photo/Ultra/Hyper Detailer Workflow
Klein Detailer Workflow
Check the note at the top of the workflow for downloading the necessary detection models and where to place them.
Krea 2 Trained Concepts
KreaLens1: soft skin, skin texture, skin pores, bad skin, innie pussy, pussy juice, wet skin, foam, oily skin, lube, sweating, goosebumps, depth of field, amateur photo, pro photo, cute, adorable, charming, girl next door, average face, cute asian, overexposure, highkey, midkey, lowkey, low contrast, flash light
KreaLife1: amateur photo, outdoor sex, party, real life, candid, candid shot, outdoor flashing, public indecency, public nudity, beach life, 3d selfie, film grain, low image quality, low contrast, flash light, blurry, soft skin, overexposure
KreaLife1/KreaLens1: 69 position, anal gape, cumshot, dildo insertion, double oral, handjob, innie pussy, object insertion, anal fingering, cuddling, cum swapping, female masturbation, finger licking, footjob, french kiss, funny tongue out, orgasm face, penis licking, pussy licking, vaginal fingering, partner fingering
SDLX Trained Concepts
Newer versions of Big Love include the concepts of older versions, e.g. Photo4 can do everything that Photo1 to Photo3 can do and more. The only exception are the Insta concepts, but they have been partially added to never versions too.
Photo6/Ultra5: ai-style, bodyscape, butt plug, candid amateur, cumshot, illustrious-style, outdoor sex, partner fingering, photoart, innie pussy, ray-style, rebellious, reflection, sensual, sexy ass, sweetheart, wedgie
Photo5/Ultra4/Hyper1: alluring, anal sex, artistic photo, artistic pinup, charming, cum portrait, cute asian, cute pinup, cute portrait, double oral, dreamy-style, face pov, fantasy pinup, fantasy-style, kai-style, lowkey, mj-style, outdoor flashing, photo art, pole dance, public indecency, public nudity, rock climbing, scify-style, sword pose, vaginal sex
Photo4/Photo4.5/Ultra2/Ultra3: 2girls kissing, acrobatic sex, average face, beautiful face, amateur pinup, anal fingering, artistic portrait, cuddling, cum swapping, doggystyle pov, even skin color, fashion photo, female masturbation, female pov, finger licking, food porn, footjob, french kiss, full nelson, funny movie, funny tongue out, insta cute, legs pov, low contrast, mirror portrait, nude pose, oral pov, orgasm face, outdoor nude, penis licking, pinup photo, pro photo, pro portrait, pussy licking, sex pov, sexy portrait, vaginal fingering, waterfall
Photo3/Lust1/Ultra1: 360 degree photo, 3d selfie, 69 position, adorable, alt beauty, anal gape, anal sex, average face, beach life, beautiful, beautiful face, candidness, cum, cute, deepthroat, dildo insertion, double exposure, enthusiastic, extreme pose, faces, fashion photo, female pov, fisting, follow me pov, girl next door, gorgeous, handjob, heroin chic, huge feet, huge hands, huge nipples, light rays, light streaks, lighting, low/mid/high contrast, lowkey/midkey/highkey, low/mid/high saturation, middle finger, motion blur, object insertion, detailed eyes, teeth, pro photo, pussy gape, real life, rooftop, round ass, sad, sexy pose, skin texture, skin tone, spread pussy, squirting, street portrait, testicle licking, testicle sucking, vaginal sex, voluminous hair, water splash
Insta1: amateur photo, fashion photo, insta selfie, pinup photo, pro photo, real life, big ass, bimbo, cute, luxurious, adorable, enthusiastic, average face, beautiful face, detailed eyes, sexy legs.
Photo2: 3d selfie, 69 position, adorable, alt beauty , amateur photo, ball licking, ball sucking, beach life, bubble butt, low/mid/high contrast, dildo insertion, double exposure, enthusiastic, extreme pose, female pov, follow me pov, forced perspective, girl next door, gorgeous, handjob, heroin chic, huge feet, huge hands, huge nipples, light rays, light streaks, lighting, middle finger, motion blur, no tan lines, object insertion, pussy gape, rooftop, sad, low/mid/high saturation, sexy pose, squirting, street portrait, water splash.
Photo1: amateur photo, pro photo, lighting, lowkey, midkey, highkey, beautiful, cute, candid, real life, average face, remarkable face, voluminous hair, skin tone, skin texture, 360 degree photo, spread pussy, deepthroat, fisting, anal gape, anal sex, vaginal sex.
Training a (character) lora
My advice is this: Take 10-30 high quality images (at least 50-100 for styles or poses), download OneTrainer, click the SDXL lora preset, choose Big Love as the base checkpoint, add your images as a concept, add tags on the Tools tab, and start training. If you train on Big Love Ultra change the training resolution to 1536 and provide 1248x1832 images.
You will only get the full Big Love quality by training on real photos or extremely photorealistic images, which is usually not the case with a fictional character. It is all about image quality and not choosing boring images. Don't add posing images to a character lora, because Big Love provides the posing later for it. Focus mainly on good portraits with a few upper and full body images. Rather fewer higher quality images than a lot of average images. Read here for traing a lora on Big Love ZT2, ZT3, Zia1, Zuna1 or Eryn1.
Fixing Anatomy
If you want to keep the seed and repair anatomical problems in an image, activate the Extra checkbox in A1111/Forge and set Variation Strength to 0.01 to 0.1. Then generate until anatomy is fine. In Comfy there are sampler nodes that support variation seed (e.g. Inspire Pack), which do the same thing. Also possible to change steps or add commas to the prompt, but I do not recommend it. Extra/Variation Seed is more powerful and convenient as it gives you endless variations with just one click. There is always a variation which looks similar or even better with great anatomy.
You don't need negative prompts for this. They are rather ineffective. You can also change the positive prompt to suppress things. Anatomy problems are also created by the prompt if you prompt contradicting poses that the model cannot combine. Better to fix the prompt first then.
Reducing Saturation, Warm Colors and Sharpness
These SDXL loras reduces the saturation and change the color temperature without reducing image quality. The lora strength defines the effect. Negative values increase saturation or add more warmth. To make an image softer use a lower strength for the DMD2 lora, e.g. 0.95, 0.9, 0.85.
Big Love Hyper/Ultra/Photo Sizes
Big Love Hyper supports these image sizes:
1664x2432 (recommend)
1536x2240
Big Love Ultra supports these image sizes:
1360x1984
1280x1872 (recommend)
Big Love Photo supports these image sizes:
1024x1496 (recommend)
832x1216
You can also use the 16:9 or 2:3 versions of these image sizes as well as landscape orientation, but the above should be more reliable.
The higher the image size, the lower the probability of good results. If you get too many problems with a higher image size, switch to the lower one. Some prompts do higher sizes fine, some do not. Do not use a lower image size than the recommended one, otherwise image quality will be lower. With higher image sizes anatomy problems occur too often.
Remove fill words and unusual tags in the prompt. They can cause a quality reduction with Big Love Ultra/Hyper as they were not trained. Shorter prompts should work better as it is less likely that bad or untrained tags are in it.
SFW Output - No Tits Please
"Is there a trick so that Big Love doesn't always generate raised shirts or visible breasts?" I get this asked so often, while it is so obvious. Big Love knows what men want so it gives it to them. Negatives don't work that well for it. So you got to speak to it in the positive prompt like you were taking to a nun. "Breasts? How dare you speak such an obscene word to me." Don't mention breasts or any of these sexy parts. Instead describe clothes a lot. Don't write down your impure thoughts. God forbid! Yeah, I know, so hard. Describe her body shape, which also implies certain tits underneath the clothes.
XL Prompting Advice
SDXL prompts & prompts from various SDXL checkpoints work with it. Pony prompts too, but they create a more photorealistic look. Natural language prompts tend to create artistic glamour images with less skin detail while normal tags do more photo-style images. Big Love XL gives individual words more attention, so you don't need to rework prompts like Big Love Pony requires sometimes.
XL1/XL2 has a tendancy towards cuteness and bright skin. If you get too much of it, remove words like cute, sweet, pale/fair skin & flash from the prompt. It also creates realistical analog photos with reduced image quality. Sometimes better to remove terms like analog or grain to increase quality. Decrease cfg to get a more natural look.
XL2.5 can sometimes get too sharp. Reduce cfg or use less steps in img2img to make images softer.
Pony Prompting Advice
Pony prompts & prompts from other checkpoints work with it. Score tags and fill words like masterpiece or perfect skin have as good as no effect usually. A negative prompt is often unnecessary unless you want to make something vanish or avoid underage, but it can also change the style of the image.
Pure booru tags and some other samplers may produce Pony style. Photographic terms or a simple prompts ensure a photo look. Sometimes it refuses to produce medium shots, portraits or close-ups unless you weigh the term heavily or remove feet, shoes, high heels, boots etc. from the prompt. wavy hair sometimes does not look so great.
If it refuses to do porn, try simplifying the prompt. If you want to do a portrait, don't describe surroundings too much. If you want to do porn, describe the action mainly and don't focus on the people or surroundings in the prompt. The largest part of the prompt should be about the main subject of the image.
If a prompt does not produce the intended image, you often only need to modify it a bit to make it work perfectly.
Using Character Loras
Big Love Pony supports Pony characters, but they look like real humans. There were some complaints about them not working, but the problem was not the model itself so far. So here are some tips:
* Try Pony as well as SDXL character loras
* Use a lora weight up to 1.5 if it does not reduce quality
* If it is a well known character, add the name in the prompt and also give it a higher weight if necessary.
* Check your negative prompt for problems. Delete it if necessary and rewrite it from scratch.
* Render a lot of images to be able to pick those with the best resemblance.
More SDXL Photorealism:
You can highly improve the images generated with this model by using img2img, upscaling, detailing etc. The SDXL Lightning lora enhances the look and constrast despite only 8 steps. Alternatively, the SDXL DMD2 lora produces a sharp, but more natural look with less contrast. Also check out the Subtle Style loras, which nicely enhance images.
Big Love License
Big Love SDXL versions are licensed under CreativeML Open RAIL++M (dated July 26, 2023), Big Love Z-Image, Ernie, Qwen & Longcat versions are licensed under Apache 2.0, Big Love Klein versions are licensed under the FLUX Non-Commercial License v2, and Krea 2 versions of Big Love are licensed under the KREA 2 Community License with the following additional terms:
For clarity, the following version tiers are defined:
Restricted versions: Ultra, Hyper, Klein, Zia, Zuna, Gwen, Cat, KreaEdit, KreaLife, KreaLens
Open versions: Pony, XL, Photo, Insta, Lust, ZT, Eryn
1. You may use Big Love without crediting the creator.
2. You may sell the images that it generates.
3. You may only run the official Big Love versions of the SubtleShader account on Civitai and Tensor Art. Running it on other public or shared servers commercially (including other Civitai and Tensor Art accounts) requires a separate license or permission from the creator.
4. You may not sell this model or merges of this model.
5. You may not merge, share merges of, or share LoRA extractions incorporating Restricted versions. You may merge and share merges of Open versions.
6. When sharing the allowed merges of Open versions, as well as finetunes of them, you must apply these same additional license terms.
7. Redistribution of Restricted versions is prohibited. They may only be obtained from official sources designated by the creator.
8. All use-based restrictions of the CreativeML Open RAIL++M license apply to SDXL versions and their derivatives. All use-based restrictions of the FLUX Non-Commercial License apply to Klein versions and their derivatives.
9. Any restrictions in these additional terms may be waived only with explicit written permission from the creator.
TLDR: This is basically the same license as before (just stated more clearly), so nothing changes for the Pony, XL, Photo, Insta, Lust, ZT and Eryn versions of Big Love. However, restrictions were added for Ultra, Hyper, Klein, Zia, Zuna, Gwen, Cat, KreaEdit, KreaLife and KreaLens versions (no merging, no redistribution). The "Share merges" condition below the license on the right hand side was only deactivated because of the restricted versions. Merging is still allowed for the open versions.
Many thanks to the creators of SDXL, Z, Qwen, Flux klein, Pony, bigAsp and Lustify for making their wonderful models available. Big thanks to RaymondLuxuryYacht for allowing me to train on his fabulous images.
Description
Please note: The KreaEdit versions of Big Love are pay-only and special license conditions apply. See description!
KreaEdit1 Big Love is a finetune of Krea 2 that can do txt2img and NSFW image editing. It is more photorealistic, a lot more NSFW-capable and has better skin texture than the original Krea 2. It produces only photorealistic images with booru tags (not anime like other Krea 2 models).
KreaEdit1 Big Love offers the best anatomy (with the least problems) of all Big Love checkpoints. It was optimized for 8 steps with er_sde sampler and beta scheduler for increased sharpness & fine detail.
Move the KreaEdit1 model file into the diffusion_model sub folder of the models folder of ComfyUI, qwen3vl_4b_bf16.safetensors (12+ GB VRAM) or qwen3vl_4b_fp8_scaled.safetensors (6-8 GB VRAM) into the text_encoder sub folder and qwen_image_vae.safetensors into the vae sub folder. To make the GGUF version work in ComfyUI, you need to unzip the second .zip file from Optional Files into the custom_nodes sub folder.
Editing component by conradlocke called Krea 2 Identity Edit, which allows character restaging, local edits, replacing a person, adding a style and combining scene & character from seperate images. Requires the Krea2Edit nodes that have to be installed manually. For more face consistency try increasing the grounding_px parameter to 1024 or up to 1536.
KreaEdit1 needs double as long to render than Klein, Gwen & Cat, but the better results with as good as no anatomy problems are worth it. You can reduce steps to 7, 6 (or even 4) to speed it up, but 8 steps is usually the best quality.
Version Recommendations
With 12+ GB VRAM I recommend using the int8 convrot version (if you have Triton installed in ComfyUI), because it works fastest on all gpus and has the highest 8-bit precision. Otherwise use mxfp8 with 5xxx gpus (if you have cuda 13 installed) or fp8 mixed in all other cases.
With 6-8 GB VRAM best grab the nvfp4 version with a 5xxx gpu (if you have cuda 13 installed), otherwise the GGUF version.
With 24 GB VRAM you can also use the bf16 version, but needs twice the generation time of int8. Only if you need absolute precision.
Lora Training
The training files of KreaEdit1 are raw versions. So train it like krea 2 raw. Different versions are needed for AI-Toolkit (single file) and OneTrainer (diffusers). OneTrainer trains double as fast as AI-Toolkit on a 5090.
FAQ
Comments (100)
Is it really worth it to upgrade from Klein 3 to Klein 4? Cause It's hard to notice any difference at all in the examples
Read the "About this version" box and draw your conclusions... I see differences in the showcase images. Klein4 definitelly produces sharper images. It received a lot of new training. Easier to do certain poses & sex scenes.
Nodes not working:
lbouaraba/comfyui-krea2edit (3)
Krea2EditGroundedEncode
Krea2EditGroundedEncode
Krea2EditModelPatch
?
Go to this link and use the installation instructions there: https://github.com/lbouaraba/comfyui-krea2edit
Last time I checked they were not in Comfy Manager. So you have to install them manually via git.
In the showcase images without any text on them there is also a txt2img workflow embedded that does not need them.
How well does your Krea2 model retain the base Krea2's pop culture and celebrity knowledge?
I did not train anything that would have erased existing knowledge. I mainly trained NSFW, lighting, posing, cuteness and skin texture.
I did not really check yet. Not really my focus. But can do if you want.
Feel free to PM me if you want to check anything in particular, I will do a quick test for you.
how do i get better results with zit,,, the vagginas upside down joysticks etc ,, or anything semi diverse it just mangles
Yep, penises are a problem with zit. Maybe there is a new lora that improves it. Did not work with zit for some time now as I focussed on other models, so also did not train such things yet with my Z models. Krea 2 definitelly has better anatomy with NSFW images.
Is krea 2 fast as zit in generations speeds, or DO i have to purchase your best zit to fix this?
@jeremypierre076910 Actually ZIT is faster with only half the params. So twice as fast. But I never released fast quant versions of my Zia & Zuna models, so practically the speed should be the same.
I did not find the time to train my Z models more on sex stuff yet (which would improve penises) as I focussed on edit models. Krea did fine penises right from the start. So much easier to train & improve compared to Z.
More Face Consistency with KreaEdit1
For more face consistency try increasing the grounding_px parameter of the Encode node to 1024 or up to 1536.
Both encode nodes or only with positive prompt (if negative is empty)?
@karstenthiel94467 No, only the positive prompt is used. With cfg 1 the negative prompt is always ignored.
pls can you upload it on tensort.art i have my credit card on it <3
Try Fanvue. I have to limit my TensorArt offers because their withdrawals fail too often & are limited. Otherwise contact me via chat for other options.
ok how fanvue works ? I have to take a 20 dollar subscription then buy the model ?
@toto47677 No, you click on Follow to see the individual models. I wrote that at the top of this page and the top of the Fanvue page hoping someone would read it. 😉
@SubtleShader And bang, right in my face.
First off, I’ve been a developer for 15 years, so my IQ’s over 100, and I’ve read the explanatory post on Fanvue.
But there seems to be a discrepancy between your explanation and reality, because I follow you on Fanvue but I can’t see the models – everything’s set to private.
That’s why I was asking whether I needed to take out a $20 subscription to access the private content and then buy the model.
@toto47677 No you don't. You just need to follow me to see & buy individual posts/models. Fanvue makes these rules and not me. I don't have any influence on that. I heard from a few people that some security software might block Fanvue features. Might be the case with you too.
@SubtleShader i don't know i have follow you but all is private :/
I already have you sdxl pro, klein 2, klein3 and gwen but not your krea. I'm sad ...
@toto47677 See my chat message for a solution.
Downloaded the KreaEdit1 model, but cannot get the otunetloaderw8a8 node to be installed. Where/how can I get it ? Google wasn't very helpful in my quest to find this on huggin somewhere.... Thanks for the amazing work btw. Klein4 is also simply amazing !
Thanks! I actually mentioned the link in the description and in the "About this version" box. But here it again:
https://github.com/lbouaraba/comfyui-krea2edit
Follow the installation instructions there. The nodes are not registered with ComfyUI Manager yet, so you need to install them manually. Afterwards restart Comfy.
@SubtleShader Did just that but it still doesn't appear as a present node... and it doesn't seem to be connected to anything when using the embeded WF with the example pictures... I must be doing something wrong... I'll keep digging... Klein4 works like a charm !
@charlesdelavigere743 Maybe you need to install the latest version of Comfy. krea 2 needs it anyway. Nothing else comes to my mind at the moment.
Since the manager button disappeared it became a bit more difficult to update comfyui (at least for me....). But I'll find how to do it.... you model deserves it !
@charlesdelavigere743 Actually I still have a Manager button. I guess, you need to reinstall the original Comfy Manager. Was part of Comfy for some time, but sounds like they removed it again.
I had to edit the config.ini. From "network_mode = public" to "network_mode = personal_cloud"
@21omen266 and this made the manager button reappear ? I tried several things lat night to get it back, but nothing successful so far... The good things with SubtleShader models is that it makes me discover new sides of complexity within ComfyUI and ways to get them fixed :-)
which comfy version do you currently have running ? I do have 0.22.3 and frontend version 1.43.18 ! Apparently version 0.27 is available...
@charlesdelavigere743 You need at least ComfyUI 0.25 to run Krea 2. You should update.
This explains why... I try to get the "manage" button back as my update.bat file doesn't want to work...and I don't feel like reinstalling the whole things, download all the dedicated nodes, etc...
@charlesdelavigere743 yeah I run 0.27. this fix of mine was necessary because it refused to install the extension at all
@21omen266 So with the change in the ini file you were able to get the manager back and do an update ? I'll need to test this once back @ home !
@charlesdelavigere743 I was already on 0.27. I don't have a Manager. I have Extentions (the same?). And there it wouldn't install the needed Nodes that caused Errors (comfyui-gguf). The ini change fixed that. Sorry if I misinterpreted your post 🫣
@21omen266 You mean update are possible through the Extention button ? Gosh, I didn't see where this was possible...
@charlesdelavigere743 NO, you can just update the nodes there. ComfyUI is another story. I run it on Linux, there are multiple steps in the terminal to update Compy.
@charlesdelavigere743 Try this https://docs.comfy.org/manager/install
Otherwise this https://github.com/comfy-org/ComfyUI-Manager
@SubtleShader I tested the github option already and it didn't make it. I'll try the other option you mentioned. I hope this will fix it and let me update the whole thing :-) I'll let you know !
Thanks.
@charlesdelavigere743 I'm not sure if this might help, but I added https://github.com/BobJohnson24/ComfyUI-INT8-Fast in addition to comfyui-krea2edit and that was the final piece that I needed.
update successful ! 0.27 installed and running ! Thanks everyone ! Now I just need to get the krea2 diffusion model (absent from the available list). I can get images but they don't look (yet) as good as the provided exemples... and the Manager button is back ! :-)
I'll need to check again once @home, but it seemed like the Otunetloaderw8a8 node was not connected to anything... it seemed strange, but as I tried a few prompts and it worked, I didn't dig more into that direction... but this seems unusual...
I only see a 17€ option. Is this supposed to be the 10$ thingy?!
Click the Follow button at the top. Fanvue only allows displaying & buying individual posts for followers. Nothing I can change about that. Sorry.
@SubtleShader I'm an idiot, sorry. Definitely not 100IQ ;-) So there i buy a picture and then i get access to KreaEdit1, right?
@21omen266 No problem. Idiots never say they are idiots (unfortunately), so you are not one 😆. You get a Fanvue DM afterwards with the download link.
@SubtleShader It worked 🥳 Thanks and have a nice Sunday!
New Training Files for KreaEdit1
New working training files are now available for AI-Toolkit and OneTrainer under Optional Files. Please delete the old diffusers zip file. It contains invalid weights that produce useless loras.
Best Krea2 model I have ever tried. And also best model I have ever tried.
Thanks. Krea 2 is the best model I have ever trained. Learns much faster & better than anything else.I hope it will set a new standard and make all other model creators train new models at least on a bit on NSFW. There is no going back.
Very Good Work this KreaEdit1. GG
Thanks a lot!
Why not have ur best models for purchase always not just for set to creation for the site?
Civitai only allows15 day periods of purchase called Early Access. 30 days or unlimited would be better. But I also offer my models on Fanvue and TensorArt (Read below the showcase images). Alternatively you can always tip me buzz on my profile page & I will send you a download link via chat. Contact me via chat for more details.
BigLoveKreaEdit1_bf16.safetensors doesn't seem to work on the Krea2_Edit workflow (the one from GitHub). It just produces those random noise artifact images. Am I using the wrong model for edit capabilities? txt2img works fine.
@SubtleShader Where do you get the base version of your KreaEdit1 bf16 model if not here?
I downloaded the bf16 version after paying the buzz. Installed the workflows/custom nodes from GitHub. Using krea2_identity_edit_v1_1.safetensors with lora weight of 1 as advised. 3:4 aspect ratio for the source pic. Steps, cfg, sampler, and scheduler as advised. The output pics are just random noise.
If I crank down the krea2_identity_edit_v1_1.safetensors lora weight to around 0.4, it sometimes manages to adhere the edit prompt but the output is still very distorted and noisy.
@elaaa Please drag & drop one of the showcase images above into Comfy. Don't use a workflow from Github. The identity lora causes the problems. Remove it.
On site generation will be available in 15 days?
If you mean KreaEdit1, Civitai will likely only support txt2img for it. The editing feature is nothing official. Requires bidding 1k buzz for it to appear on the generator.
@SubtleShader So, in 15 days I'll only be able to download the model and that's it?
@nerdoo That is just what the Civitai standard message says that I have no influence on. In the "About this version" box it says: "Please note: The KreaEdit versions of Big Love are pay-only and special license conditions apply. See description!" I might make it freely downloadable at some point in the future, but no idea when.
Making it freely downloadable after 14 days means throwing away up to 50% of the money. I'm doing this full time as my main job and have to support a family. Would likely have to do another job otherwise & not release any new models anymore as it would not be worth the effort.
@SubtleShader ok, I understand, thanks for the explanation.
pls upload krea2 on tensor. love u
It is on TensortArt now too
Dual references image possible ? Can you provide two references images for krea2 as it was the case for klein4 ? It would be great if this was possible...
Yes, possible. The conditioning node has a second image input.
@SubtleShader Ok, I'll tweak this around... Thanks ! I also notice that the krea2 seems to keep a lot of information from the ref image. For instance if I provide a picture of a person wearing a coat and ask for a picture with, I don't know, let's say, a tropical short... the coat will stay there most of the time... I tried to prompt with "no coat", "remove coat", "no extra clothes",.. but it didn't make it most of the time... Any special trick when you want to remove something ? Same issue if the ref image shows someone with a necklace or earings, these will stay in the final image.
@charlesdelavigere743 The edit functionality is rather new, so does not work perfectly yet. Will certainly improve over time. You could try to remove the necklace, coat etc. first,.then use that as a reference image. You could use Klein or Gwen for that if KreaEdit does not do it efficiently.
@SubtleShader This is exactly what I was thinking about. I'll pre-process the image with Klein, then use it as a "new" reference with Krea... will test that @home later !
The resulting images with Krea are simply amazing !
Can you do a comparison table of your models to help decide which one to pick? I always thought Klein was the best but seeing you release stuff for Krea and Cat makes me wonder those could be better or maybe they are for different purposes. Thanks!
Each edit model has its strength & weaknesses:
Klein: Versatile editing, great skin texture, a lot of bad anatomy with more advanced posing
Gwen: Great at posing & prompt following, softer results, sex scenes are ok
Cat: Most reliable at reproducing faces, sometimes lower image quality, worst at sex scenes
KreaEdit: Great image quality, best anatomy & NSFW results, editing not always effective
To compensate you sometimes need to use more than one model. I always try to train the things that a model is less good at (if possible) to improve it.
Big Love Eryn1 is freely downloadable. Enjoy it!
No it's not. It's 16k lol
@roboticone Very funny. Best joke ever. 🤣
I tried Cat1 but can’t get a single image to maintain facial consistency with input image, while Klein works way better…using example workflow, strange. Not sure what I’m doing wrong
Might be your settings..My tests indicated the opposite. I'm always testing with up to 5k images. If you check out my showcase images, you can see how well Cat1 does it.
Fanvue Subscription Problem
Fanvue blocked my product delivery method. It's not about NSFW, but rather silly. If you subscribed on Fanvue please contact me via chat so I can continue delivering new releases to you. Sorry for the inconveniance!
Your model is incredibly effective. Is it possible to do the same thing but with your ZImageTurbo models? If so, can you tell me which nodes to use in ComfyUI? I can't seem to set up a workflow that works.
You mean KreaEdit1 is effective? You mean editing with Zuna/Zia? Sorry, not possible. Editing functionality would need to be trained on Z first.
@SubtleShader Yes I mean KreaEdit1. But want to use image reference on other models like ZImageTurbo if possible
@Foreign77 May come in future. I'm collaborating with the guy who trained the editing functionality for Krea 2. Z would definitelly be the next architecture to train it for. Maybe I will do it myself in future, but it is quite some work & effort. Don't expect it very soon.
@SubtleShader Nice to read that. Take your time
Love the anatomy/skin quality on KreaEdit1 but I can't get the face to stay consistent on edits even with your exact recommended settings and grounding_px up to 1536 any tips on what might be off?
1536 grounding might be too much. Also takes longer to calculate. I mostly use 1024 or 768.
The Krea2Edit nodes were updated. Now there is a new Match Model node (add it before the sampler) that improves face consistency with the ref_boost param of 4. Reduce the value if it just copies the face without modification too often.
@SubtleShader Just tried it and now is much betteeeeer : ref_boost 4 + grounding_px 1024 fixed the face consistency issue completely. Great work on this, thanks so much for the help, really appreciate it! ♡
@SubtleShader This Tip helped out a lot, was increasing the grounding_px value to max of 1536 and it was helping with face consistency but still too far from starting image and it was also doing weird stuff like adding cracks and other odd things but the ref_boost to 4 & grounding_px to 1280 was just right.
I am having real trouble getting Krea 2 edit to remove clothing from an image, It refuses nudity without extra loras.
It does not work always right away. Try more seeds and also adjust the prompt. It should also help to use a headshot without any clothes, so it cannot pick them up. So just crop the image.
I advice against using extra loras. They do not influence the editing itself and may reduce image quality. I plan to do some editing training myself on top to improve NSFW behavior. The original editing training did not include NSFW.
@SubtleShader Version 1.2 is far better. I think you need to update your model with the new Edit function. I bought your version, but when they released Edit 1.2, I switched to another checkpoint.
@rejariel399 My extensive tests show that v1 is still better than v1.2. The newer version was trained with lower resolution and material that did not include face replication.
The updated Krea2Edit nodes improve it with any version, more than any training could. If you want to use v1.2 with KreaEdit1, just remove v1 from KreaEdit1 with -1.0 strength and add v1.2 with 1.0 strength.
With NSFW worflow it works fine. With NSFW loras even better.
Bias to poses or clothes from reference its base Krea 2 behaviour.
A bit weird behaviour of Krea 2 Edit - its actually not working with just face references. To achieve likeness it require portrait reference not just face.
is your Biglove version based on Krea 2 turbo or Raw? I ask so that I know what cfg / steps combo I can push.
It is turbo but I also provide raw versions without turbo and editing. You can use the raw version for generating images too.
@SubtleShader Oh awesome, is it the file listed under "Optional Files" BigLoveKreaEdit1_raw_OneTrainer .zip?
@Shuttergrenade No, it is BigLoveKreaEdit1_raw_AI-Toolkit.zip. The OneTrainer one is diffusers format.
@SubtleShader I have downloaded the Raw KreaEdit1 zip file 3 times between two computers but when uncompressing using 7z it says headers are missing and stops extraction. :( could you re-upload pls. I have had great results editing with your KreaEdit but I would like the images to be larger than 1.5MP without the weird distortions start appearing when editing at bigger sizes and from my understanding I would need your raw model for that, correct?
Hi, I’ve purchased quite a few of your models and I’m using them for subtle fine-art nude / glamour photography rather than explicit content.
So far I’ve had the most consistent success with ZIA1, but I’d really value your opinion on which of your models you’d recommend as the best main production model for this kind of work.
My biggest priorities are:
- strong prompt adherence
- good lighting control
- strong photographic/artistic style
- photorealistic results
- reliable anatomy and faces over larger batches
- tasteful fine-art / glamour / boudoir output rather than explicit content
Would you still recommend ZIA1 for this, or is there another model of yours that you think would better suit this workflow?
Thanks
Yeah, Zia1 is great for your use case. Currently my favorite model architecture is Krea 2. Best learner of all models that I trained on. Also best NSFW anatomy because pretrained on it. Z also has great anatomy for non-explicit NSFW, but less for explicit stuff. I still plan to train Zia & Zuna more on explicit stuff but other models distracted me so far.
So I would recommend you to try KreaEdit1. The edit functionality is also something that Z does not offer yet. I will soon release a more candid version of krea, but I guess not what you look for.