Big Love Klein3 is also offered for $10 on Tensor Art. You can get Big Love Eryn1, Ultra5, Photo6, Zuna1 or Zia1 for $5.50 and Gwen1 or Hyper1 for $10 on Fanvue. Click 'Follow', buy a post, check your Fanvue DMs! A subscription includes the latest release and new ones within 30 days. Any problem? 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.
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.
Settings are 4 steps (2 steps for upscaling), cfg1, Euler sampler, Beta scheduler, 832x1216, 1024x1536, 1280x1920 or 1536x2240 pixels
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 less reliable, but generates nice results too.
Big Love Gwen1 allows commercial usage without additional fees (if no auto-generation involved) unlike Big Love Klein. With its better anatomy it is easier to get good editing results. As fast at 6 steps as Klein at 4 steps. Check the "About this version" box on the right for info on the individual variants.
It is recommended to use ComfyUI. The Gwen1 workflows offer a skin enhance feature that is not available in Forge Neo. Move the Gwen1 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 for Gwen1.
Settings are 6 steps (3 steps for upscaling), cfg1, Euler sampler, Beta scheduler, 832x1216 or 1024x1536 pixels.
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.
Big Love Hyper supports generating at image sizes of 1664x2232 (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.
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.
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.
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.
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.
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.
ComfyUI Workflow:
Download one of the following images and drag'n'drop it onto Comfy:
Klein, Eryn and Gwen workflows are embedded in the showcase images.
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
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, the Big Love Z-Image, Ernie & Qwen versions are licensed under Apache 2.0, and the Big Love Klein versions are licensed under the FLUX Non-Commercial License with the following additional terms:
For clarity, the following version tiers are defined:
Restricted versions: Ultra, Hyper, Klein, Zia, Zuna, Eryn, Gwen
Open versions: Pony, XL, Photo, Insta, Lust, ZT
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 and ZT versions of Big Love. However, restrictions were added for Ultra, Hyper, Klein, Zia, Zuna, Eryn and Gwen 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
More photorealistic, improved lighting, less saturation, darker
FAQ
Comments (118)
is there a trick so that you don't always generate raised shirts or visible breasts?
Just add breasts or nude breasts. Or shirt pulled-up.
You can try this nsfw embedding, it has helped with other checkpoints but i haven't tried on BL yet https://civitai.com/models/1302719?modelVersionId=1601074
@ElectricDreams Would not recommend it. Was done for Illustrious. Big Love has zero Illustrious. Is also negative, so won't work with DMD2.
cleavage might work
@SubtleShader that's exactly what I don't want
@SubtleShader Why dont work negative with dmd2? Not just 1 cfg exist. The steps can be higher too. Its not set in stone, how to use it.
Anyway dmd2 has problem with clothes so much, i gave up to use.
@leojefferies306 Yes, negatives work with cfg 1.25 or higher, but I don't use negatives anyway. Who needs clothes on women?
@serendipity4k545 Sorry, was thinking of what I want and overread the "don't". Guess, I need to put this in the description, because I get it 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 a 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! Describe her body shape, which also implies certain tits hidden underneath the clothes.
@leojefferies306 Read above!
@SubtleShader Its the shape of clothes, where dmd plays dumb.
@leojefferies306 I guess, DMD2 was not trained too much on clothes.
@SubtleShader Yeah, poses are perfect.
Thanks for this update! It means I have to update my models as wel...
No need. Still based on XL2 with some loras added and sharpness improved. If you trained on XL2 or XL2.5, they should be fine.
@SubtleShader Did you just take away my need and desire to train? :)
@GavinStormar Feel free to train the shit out of your GPU 😂!
@SubtleShader Hahaha. LOL
@SubtleShader this is amazing! I was concerned about this too lol. Thank you so much!
I have a question:
I wanna use my own Character LORAs (that I trained on this checkpoint btw) on this checkpoint. I know according to your description that in order to get the best results I need to use Dmd2 with very low guidance 1-2 (and LCM). However, that CFG is way too low for my Loras.
Is there a way around this? How does one use character loras with this checkpoint if the optimal CFG is 1-3?
Am I just going to have to accept that I won't be able to use the optimal settings when incorporating characters loras?
You don't need to use DMD2. Just compare and see if it works better with or without it. Loras don't have an optimal cfg. They are just merged with checkpoints on the fly. The cfg depends on the checkpoint and special loras like DMD2.
Use aDetailer with the lora trigger and loras in the position prompt, including the DMD2 lora.
I have no trouble using my character Lora's even the ones I didn't train on this checkpoint maybe something wrong with your lora?
@MyaSturbate Yes you are correct, I did something wrong in the training. I don't know what it was but when I retrained it and plugged it into Big Love with dmd2 it worked perfectly. Wish I knew what exactly I did wrong before.
Wow, the pony model is exceptional with LCM and 1536 results are something to behold.
You mean Big Love Pony2 images are much better with DMD2 & UltraRes lora? Did not try yet. Can you post something?
Pony versions don't have anywhere near the same level of variety / concepts as the SDXL, I also don't think it comes close for photo realism.
@ronikush Pony2 of Big Love can created some nice photorealistic images with certain prompts that suppress the Pony components, but usually there is still a final touch missing.
When I see an update of my favorite model, I'm always thrilled!!! thank you!
The best model ever made just keeps getting better... You are the checkpoint fkn maaaster! -
One thing I have noticed though 1000's of generations at this point, is that with DMD2 lora a lot of variety / concepts / styles are supressed. This is not a fault with Big Love though, This happens with DMD2 lora on every model... Not found any way round it other than finding or creating loras to fill in what goes missing.
Thanks. I guess the high speed generation comes at price. My main reason for liking DMD2 is that its results look closest to real photos than anything else I have seen. People often forget that photos are often less perfect than what SDXL can generate. SDXL produces a rather painting-like style by default.
Retraing DMD2 is out of reach for me because requires huge ressources. I tried modifying DMD2 but no luck so far. So I'm trying to optimized Big Love for DMD2.
What concepts & styles went missing? Let me know.
use dmd2 as refiner if you want to keep all variety / styles and add photorealism
So I trained a character LORA on this checkpoint and I want to inpaint that character on an image. Has anyone found a good inpainting checkpoint that works decently well with Big Love-trained LORAs?
Inpainting checkpoints were a thing with SD1.5. Not needed for SDXL. Just use Big Love. If you want to use another checkpoint, train on that checkpoint, not Big Love.
AMAZING!!!!!! SOOOO GOOOD CHECKPOINT
Does anyone know a penis slider LORA cause the penises here are small. I tried huge penis and large penis but its still small
Big Love keeps it real 😉. Tried the Civitai search?
There are fuckton of elephant penis lora.
@ronaldtodd961 On sdxl 1.0 no
@SubtleShader Yea I tried searching SDXL loras but none came up. Pony has some but Pony loras don't seem to work on sdxl either
@Yourmomd https://civitai.com/models/424318/penistest
maybe
@ronaldtodd961 unfortunately it is still the same
@Yourmomd Pony loras work. At least with Big Love. Just try some.
This gives some beautiful results. Made me come back to SDXL after months out. And it can do male really well if it is a one subject only.
But It's pretty much impossible to get any gay (non vaginal) male sex with it. I tried stacking multiple loras and it won't work, unfortunately.
Interesting. Never tested that. I guess the merged models were never trained on it either. Might work if you train a lora with Big Love as the base model. But not sure.
@SubtleShader I know it had some pony I guess, that does know anal gay sex, but the other probably not. It's very biased towards straight sex to the point not even 5 stacked sdxl loras worked. Maybe training on it, but, well I don't have the dataset, and I'm more focused on flux nowadays. But it's a nice model anyway
Hey can we agree, v3 is the best biglove model.
i mean holyfuck combined with faceIDv2 sdxl i can achieve photorealistic image with near 100% Resemblance with the subject face.
Would not have released it if it was not better.
do you have an example workflow with faceIDv2 ?
@mfudi You only need to google it. Plenty to be found.
I think yours is the most realistic Pony based model and your workflow is great too.
Sorry if I skipped over it, I was skimming through the description but didn't find any information. What's new with XL3?
There is a "About this version" box on the right side above the description. To quote: "More photorealistic, improved lighting, less saturation, darker"
@SubtleShader Sorry, I didn't know about this feature. Thanks for model, and your patience!
Best SDXL NSFW model
Hmm, it is great at prompt inherence but as far as if one is going for photo realism with NSFW I'd disagree, it seems to have this overburnt/artificial look to it that is inescapable. BigASP, Lustify and Pornpro(amatuer/XL) all do NSFW very well and realism very well too.
Of course if realism isn't a concern carry on but just chiming in for anyone who is looking at this and thinking it is the best.. it's up there but certainly not unequivocally the best
I'm confused. Is this Big Love Pony2 based on SD 1.5 or SDXL or Pony?
I want to train lora based on Big Love Pony2 and I don't know what resolution the photo should have for training based on Big Love Pony2.
Big Love is SDXL. Pony is also SDXL. Only the upcoming Pony v7 won't be SDXL anymore.
@kareslav Ideally, images should be 1024 pixel in height or width minimum. I usually delete any training image that is less than 900 pixels. The bigger, the better. The training software will automatically downscale to the right size.
@SubtleShader Thanks for responding, buddy! I tested about a dozen different models and big love pony2 was my favorite. Thanks again for all your hard work!
@SubtleShader So if I train lora on 1024*1024 images, the generation results will be better than lora trained on 512*512 resolution images?
@kareslav SDXL is trained on 1024 pixel minimum. Much less and you degrade quality. Better train on 2:3 ratio like 1024x1536 or 675x1024. Very few people render 1:1 sizes like 1024x1024.
@SubtleShader Thanks for this advice! Because I asked the same question to civitai bot and he told me that in my case it is absolutely necessary to train lore on 512*512 images :)
@SubtleShader I tought 4k is minimum quality.
@timthomas131 Default SDXL training settings is 1024 pixels
@SubtleShader I meant about dataset. So fullhd enough.
@timthomas131 The training software scales it down to the required size.
nice
Certain loras are working better for me in XL 2 than XL 3 but over all I think ALL the Big Love SDXL models are great. Thanks!
Yes, lora fidelity is better in 2.5. I think the reason might be the merge of subtle lightning in BL3 that affects lora fidelity, even at very low weights, based in my own experience.
@ElectricDreams I think he means XL2 and not XL2.5. I departed from a pure merge in XL2.5 by enhancing sharpness and in XL3 by making it darker & improving lighting. But I still like the changes. Feels more photorealistic.
@SubtleShader then i understand even less, because 2.5 is much better than 2 (without any trade offs).
Guess I should have given more info. I was using Big love XL 2 in tungsten.run and I went to go re-run a job/image that was in my queue and noticed there was now XL 3 version so i switched the model to that and re-ran the job. I don't think it liked the combination of LORAs. I was using a bunch of them, the ones that give the women more size and cellulite. Anyway, when I tried to run the same thing with XL 3 it was all blotchy with spots everywhere. That's what I was referring to but maybe I jumped the gun because it seems it was only that combination of LORAs giving me an issue because everything else I tried with 3 was better than 2, including skin color and texture, especially when trying to make darker skin. I also don't seem to need as many LORAs with version 3 to get it to do what I want it to do so my first post is probably not relevant anymore. I've done lots of playing around with it since my first post and its working really nice since.
Am I wrong or XL 2/2.5 have more skin details while XL 3 has a more flux type of skin
Have not seen any Flux-type skin in XL3, but XL2.5 seems to be a bit more sharp. But.of course, depends on what settings & tools you use.
Personal opinion not applicable for everyone: I think your best model so far are still XL1 and Pony 2. Newer mix with LustifyOLT (to much syntetic - Endgame was better) and adding DMD2 are not helping to your newer model. I am using XL1/Pony 2 mainly and comparing with every other models out there and I am still coming back to those (check how I prompt or use workflows). Nevertheless I appreciate your effort and dedication. Consider it just a feedback.
There is no LustifyOLT in any Big Love version. Did not like the results when adding it. There is also no DMD2 in any Big Love version. It is an optional lora. You can also use Lightning with newer Big Love versions. Or no speed lora at all.
I know that some people prefer older Big Love version. That is fine. Tastes are different.
After XL1 the models seem overbaked. Maybe thatswhy you love the first.
@jamescunningham530 What you call overbaked is just using the DMD2 lora instead of the Lighning lora. You can produce a similar look to XL1 with XL2/3 by using Lightning.
I have to agree, Big Love models are all I use, best checkpoints ever made! and I've generated 1000's upon 1000's of images at this point, XL1 with no DMD2 or Lighting (all of which kill off so many niche / content / variety etc of Big Love) - XL1 is king of realistic images, colours, contrast, brightness, balance, bias, its behaviour with loras, every possible metric is absolute perfection.
@ronikush I hate dmd for bad clothes only. Of course it cant generate good monster girls and such too.
What niche do you mean?
Wow. I'm not alone. I think Endgame is the best Lustify too. Yeah synthetic data is problematic.
@ElectricDreams The applied style loras are synthetic.
Agreed x1 is the best version by far.
@MILOPS Different people say that each version is the best. A matter of taste.
@SubtleShader Certainly. Not attenmpting to down play your work. From my own generation x1 has hit the spot. Do you mind if I directly message you? I have a specific question regarding it and a particular face it does so well.
@MILOPS Sure, contact me.
@SubtleShader Asked you the face question on discord. Again, appreciate the answer, explains why I like xl1 so much :heart:
@MILOPS The face was generated with Pony2 not XL1. 😆
@SubtleShader !! Hah!
I updated the Big Love description above and repleaced 1024x1536 with 1024x1496. This image size avoids the elongated body problem mostly. I recommend using 1024x1496.
Using the "DMD2 Workflow (recommended)" from the description with XL3. The negative prompt has no effect on the generated output. If I keep everything else the same (positive prompt, seed etc.) I can add anything to the negative prompt and I get no change.
Negatives don't work with cfg 1. You have to increase cfg e.g. to 1.25 or 1.5 to make them work again. I don't use any negatives. I rather adjust the positive prompt. Works better most of the time.
Why is this so fast? With 1 LoRA and 13.2 GB on a 2080ti (11GB) 1.02 s/it!
DMD2 is a speed lora. The LCM sampler is fast too.
Is this Dreambooth'able? Not super experienced wwith Generative stuff so I am not even sure Dreambooth is still the latest technique. But can I use this with little non nsfw image training to create a photorealistic image
You mean train Big Love further? Yes, possible. Easier to train a Lora or Dora. Doras are similar to fine-tuning a model. You can also prompt SFW images. No need to prompt.
@SubtleShader I meant can I train this model to generate pictures of a specific person.
@SubtleShader Hey im new with this stuff! First i only know about kohya and it doesnt have the dora option, what should i use. also i have a weird issue with the fine tuning of a model so i train it with my person image set but it seems to limits the position, situtation, background etc options when i train it. Im wondering if it s just a impression or it s a fact that training VS finetuning not add to the checkpoin model but reduce it to my persone image set. thanks!! :D
What do you think, will we get a porn model in the future, which has not strict anime database?
Pony7 knows realism out of box(good enough for better finetunes, than v6, but the future seems boring.
Since Civitai removed every possible loras, models, which connect to real person, my browsing time shortened to minimal.
Its so boring, everything is boring on Civ now.
XL is dead, barely new models, Pony is dead, barely new realistic models, Illustrious lives, but all realistic merges suck.
Fuck these faqs, this comment is boring too.
I don't know how you define porn, but Big Love can to porn images too. Some sex stuff is a bit difficult, why I created the Subtle Poses loras, which will be merged into it in future. Of course, it cannot do some of the exaggerated unrealistic sex stuff that Pony can do.
I'm not really interested in celeb loras, but easy to train one yourself. I'm not bored by fictional photorealistic characters. Most checkpoint can also do celebs without a lora, but often not as good as loras. Some better some less.
@SubtleShader I browse everyday new loras. Since the new policy, there is barely any realistic character and there isnt siple any realistic helper lora. Before this, there were this type of loras on daily basis.
Not XL is dead, but Civitai.
Two upcoming models, what is interesting only is Pony 7(easier finetune for realism, at least XL level and with the benefit of easy Pony porn) and Chroma(flux finetune for porn, which is now a generalistic model with anime and realism 50-50). Both are slow and need finetunes.
Flux guys bet on Chroma, Pony guys on pony7. But its sure, when they released, they would be just a base model(so sucks).
@marteseyousif I still bet on SDXL. Nothing better available yet.
I love this model. Imo the best XL so far.
Question:
I noticed that using DMD2 model faces tend to look less varied and the result are the typical AI woman generic face. That's because the quick steps?
Because when you use models on his full potential ther'es a lot of differents kind of faces and expresions.
That's the way it is, there's no chance or any trick to get varied faces with DMD2?
I get varied enough faces with DMD2. I like the faces with DMD2 better. But you can try to use one of my loras like Cuteness, Candidness, Beauty etc. A lot of new faces in there. Another trick is to add a made-up woman name (including family name) in the prompt as that gives you more face variation. I also often add cute to the prompt.
any advise on how i can get a fenboy/shemale im trid and tried. prompts and loras dont help, any help would be amazing , thank you
Never tried. Someone commented that if started to work with Big Love XL2, I think. He did not specify what he prompted. I only observed that a penis can get attached to a female looking person randomly if penis is in the prompt without 1boy of man being mentioned. I'm also seeing testicles turning into pussies occasionally.
@SubtleShader okie dokie i primarily use big love and just wondered if it was possible because iv tried and just can't seem to get it to produce it unfortunately
@djacid Check this out: https://civitai.com/models/1089103/tgirls-realistic Ask the creator. He can certainly help you better than me.
What I love about this model is that every face generated is unique, just like in the real world. Other models feel as if the faces generated are all very similar and I'm being robbed of the full capability of the training data. Biglove feels as if I'm more in control of what's being created. The style and look of the whole image hasn't been predetermined by someone else. Thank you (edit: I'm using it raw, w/o any of the fast rendering loras btw)
Illustrious please!
Will likely do a Illustrious version. But first a Pony3 and XL4 Version. Pony3 will contain a bit of Illustrious.
@SubtleShader Good luck with that. Finetuned Illustrious realism is nowhere near to Pony.
There are photorealistic merges, and what a surprise, they dont know anything about Illustrious prompting.
The Lustify guy made a perfect finetune and it knows everything, but, if you want to use characters, and character loras, its max semireal.
Nothing I do can get rid off extremely bad and distorted physiques,limbs and so on. No matter what settings or Lora I use, using ForgeUI...
I get it rather seldom. Some prompts with contradicting descriptions can cause it. Some loras too. A too big scaling factor in img2img or hires fix too.
@SubtleShader Thanks, got that fixed up mostly, but now the issue is way too oversaturated images with blown up exposures... Seems like a real good checkpoint, just wish I could make it work like everyone else lol.
@raAI_ If you use DMD2 set cfg to 1. You can prompt against overexposure & saturation. The upcoming Pony3 & XL4 version won't have this problem.