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
Feels more realistic than Photo2. Improved eye details. Tends to do bigger boobs, asses and lips. Less exposed breasts and pussies if not prompted. Trained on 2075 images. A bit of Lustify 7 injected. Does Instagram-style poses & images as well as NSFW as based on Photo2. 16 concepts were trained: 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.
FAQ
Comments (123)
is photo 1 friendly with the loras xl of celebrities? and what about Subtle Lighting from BIG LOVE UNDERWATER XL3 & Subtle Lighting from BIG LOVE UNDERWATER XL3?
I'm not into celeb, so please test yourself. Try an older version of Big Love, if there are problems. Lighting was trained into Photo1, so no need for the lora. Underwater was not.
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, 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.
Hmm. With version 2, I'm getting very poor results with recommended settings and none of the loras seem to work.
Photo2? XL2? What settings? Do you use the DMD2 lora? What loras don't work? Please more details.
Does Photo2 do penis licking?
Yes
@SubtleShader How do I trigger it? Also been trying to trigger "middle finger" but results aren't consistent - other fingers keep going up
@qustionauthrity838 If licking penis does produce results, also add 1boy, oral sex, fellatio to the prompt. Yes, middle finger needs too many attempts to work. Will train again in Photo3.
@SubtleShader Squirting also seems hard to trigger - beeng using "squirting" with no effect
@qustionauthrity838 Yep. Had to do some merging to improve image quality which weakened some concepts. Image quality has higher priority. Will try to avoid that with Photo3 and retrain everything again. Squirting works a bit better with water ray or similar additional tags, but not as good as before the merging.
i am always looking forward to any updates on these models, but I always end up coming back to sdxl1 model of yours, it just better image clarity wise imo. I also often use pony3 model, as it is super awesome for image understanding but I find it rather biased towards brown skin? Could just be me.
I’d love to see a model of these two combined, the image clarity and character understanding from sdxl1 and the prompt understanding from pony3. I have tried to merge the two together, but i did not succeed in any very good results, only mediocre
Ps: this is not a bad review or an insult in any way, but rather a suggestion or at least a heads up that some of us still prefer your sdxl1 model
XL1 is less realistic than the newer models (except maybe Pony3). Pony always reduces realism. XL1 still got too much of it. Sure, if you like the Pony-like look better.
@SubtleShader Yes, but girls are much more beautiful, loras works better, celebs are better and dont have 90's porn magazine look.
@mccabeadam592 Still less photorealistic. Better is very subjective.
@SubtleShader Yes, of course, its less realistic.
I don't want to oversell this checkpoint but Insta1 totally crossed the Rubicon of excellence on SDXL. A point of no return in the evolution of generative AI. This is the most transcendetal checkpoint in a long time. Truly the Juggernaut of Porn. Nothing will be the same after Insta1.
i dunno, i do test sever checkpoints, never could find any love for big love. It kinda looks overcontrated/sharpened/staurated to me, the realism is quite okayish tho.
@NeuroFunkeR That is what people want, otherwise not such high download numbers. You can reduce it with certain settings and prompting during generation. And you don't need Photoshop to edit faded images as is the case with many other checkpoints. It is meant for fast generation without additional fuzz.
@SubtleShader thanks for reply, also was not a hate:), if you could tell i few settings/promts how to make to looks more subtile/RAW, it would be appriciated and i'll try it again :-)
@NeuroFunkeR it was a pretty legit opinion tbh. Nobody thought that was hate. Now that you mention it, it would be interesting to see how BL would look with mored toned colors. Mmmm... Interesting.
@NeuroFunkeR You can reduce steps, lower the DMD2 strength a bit for more softeness. Prompt for neutral skin tone, low contrast, low satruation, cloudy. Check that your prompt does not contain anything that hints at colorfulness. Also use lower settings when upscaling. Photo2 has a bit lower saturation than the others. I plan to train low contrast and low saturation more in future versions.
@ElectricDreams The initial beta version had rather moderate saturation. Did not like it. It did not look good to me.
FP32?
That better? I only trained bf16, so releasing a fp32 version makes no sense for existing checkpoints.
@SubtleShader ОК
insta1 it's very unstable for me. sometimes the quality is amazing, and sometimes the upscale effect is very noticeable. or is it just me but I also have big problems with hands. I didn't see this in the previous model
I did not notice more problems with hands than in previous versions. Try different upscale settings. Sometimes they need to be lower than txt2img settings. Also reduce denoise. I often use 0.3 or 0.35 as 0.4 can be too much sometimes. Upscaling can add smoothness as the original txt2img result is already very good. Best do txt2img at 1024x1496 for optimal quality.
@SubtleShader When you released the photo1 model, I was just learning how to use comfy UI and how to work with nodes correctly. I'll try everything.
Hey, BigLove2 looks amazing! - I just went to try it out in A1111 and the images are super blurry though - to the point where they're all basically grey and impossible to make out. It's like looking at the images through heavily frosted glass - and idea what I'm doing wrong? - I'd love to get this model working.
First, check settings. Are you using the dmd2 speed LORA with CFG set to 1.0, Sampler to LCM and Scheduler to Exponential, steps 6-8? Big Love checkpoints are optimized for use with dmd2. The way you describe it, it sounds like you may have the right settings, but have not added the dmd2 LoRA? If that was not it, reply again with more context, please.
You need to use the DMD2 lora with LCM Exponential, 8 steps and cfg1. Download it here:
https://civitai.com/models/1608870/dmd2-speed-lora-sdxl-pony-illustrious
and place it in the A1111 /models/lora/ sub folder.
Then add this to the prompt: <lora:dmd2_sdxl_4step_lora_fp16:1.0>
@SubtleShader Awesome. Thanks so much!! It works now.
What exactly are the differences between Insta and Photo1/2? If I produce images with the same exact seed and settings in both checkpoints, what noticeable differences will I notice? Not criticizing, just curious.
Quoting the above "About this version":
"Feels more realistic than Photo2. Improved eye details. Tends to do bigger boobs, asses and lips. Less exposed breasts and pussies if not prompted. Trained on 2075 images. A bit of Lustify 7 injected. Does Instagram-style poses & images as well as NSFW as based on Photo2. 16 concepts were trained: 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."
Photo 2 is great. Thank you.
photo 2 is lot more responsive to concepts now.
Thanks. Will try to make it even better with Photo3.
@SubtleShader can you add full nelson concepts to the photo3.
Thanks for the new release. What is Insta about? What is the difference with Photo?
Quoting the above "About this version":
"Feels more realistic than Photo2. Improved eye details. Tends to do bigger boobs, asses and lips. Less exposed breasts and pussies if not prompted. Trained on 2075 images. A bit of Lustify 7 injected. Does Instagram-style poses & images as well as NSFW as based on Photo2. 16 concepts were trained: 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."
I added the DMD2 lora and use it in A1111, but it looks all weird. Im likely doing something wrong.
The model itself looks amazing so i gotta figure out how to use this DMD2 stuff. So many models have it these days...
Just add the dmd2 Lora and switch to LCM and Exponential and only use 8 steps and cfg 1 (if you need negative prompt can go up to like 1.5ish usually )
You need to use the DMD2 lora with LCM Exponential, 8 steps and cfg1. Download it here:
https://civitai.com/models/1608870/dmd2-speed-lora-sdxl-pony-illustrious
and place it in the A1111 /models/lora/ sub folder.
Then add this to the prompt: <lora:dmd2_sdxl_4step_lora_fp16:1.0>
@DrAIvyn cfg 1.5 is too much. I would only recommend cfg 1 with Photo & Insta. I did not use the negative prompt in over a year & I am fine. It was great to improve quality in SD1.5, but is pretty useless with SDXL. You can always adjust the positive prompt to fix things.
@SubtleShader I'm not sure I've used higher in a long time either so I relied on my photographic memory which said 1.5 was okay, but you are correct, of course. Think I had used like maybe 1.2 when I used to feel like i needed a negative.
@DrAIvyn cfg 1.5 might work with XL but not with Photo/Insta. Will retrain everything for Photo3 and a new training strategy so hopefully cfg 1.5 will be possible again.
@SubtleShader Ah LCM, that must have been why. I will try this when I get home!
can you add full nelson concepts to the photo3.
Yep, will try. If not, then in Photo4.
Did you alway use Anteros? Aint know man, that is a pile of yellow crap plus ugly as hell.
I merged Anteros in XL2. But very low ratio. Does not really come out in Big Love. It is buried under many other models. Plus there is a lot trained on top with the Photo/Insta versions.
@SubtleShader What models are exatly in the new versions? Beside your own training set.
@rodnernoland919 I only merged loras in XL2.5 to XL4. Only training since Photo1. Apart from a bit of Lustify7 in Insta1 and a bit of Illustrious in Pony3. I stopped merging checkpoints after XL2, because that did not improve anything. Why I had to start training.
What??? Insta1?
Bro, chill, I haven’t even survived photo 2 yet 💀
Haha... feel free to ignore newer models till you are through with an old one.
can photo3 include lactating breasts?
I'm sure there is a lora for that. Don't want to add too niche stuff.
So many requests for concepts, it will slowly become a new base model.
Your Photo models are fast and good, but Pony 3 and XL 4 are superb, not as fast than Photo but the pictures are fantastic. Thanks again for giving us the choice. These models are amongst the best of the best.
Thanks for the compliment. There is no speed difference when using the same settings.
@SubtleShader OK, I've used DMD2 with Photo and more steps for the others. Good to know, thank you, I will try several settings. Again, great models. Thanks.
Edit: After some advice from Subtle, and following his instructions exactly as listed for "Best Quality", the model is working really well for me. It understands concepts that not many other models do, and it's wicked fast.
Struggling with high contrast in Photo2, especially in hair. Things seem too sharp. Although, I'll admit I didn't follow your settings to a T. (Which, if that's the cause, I take 100% fault for and should not reflect poorly on the model!)
Using Forge, DMD2 Lora, 8 steps base, 1cfg, lcm exponential. Hires fix 3 steps (tried 8, but it over baked), cfg 1, 1.5x. I know you mentioned using Img2Img to achieve better results, but unfortunately that doesn't work with my workflow so I've been using Txt2Img Hires Fix.
I've tried playing with some other settings, but the contrast issue seems to persist to some degree.
Any other tips? I love the breadth of ability this model has, and really want to make it work.
There is no good hires settings for dmd models. Only 1 setting(euler a karras), which dont change the face, but dont fix the eyes also, and make the skin artifacty.
So, you use other upscale method or 3.rd party software.
Don't use hires fix. Low quality and it forces you to use low hires steps with DMD2, so quality is lowered further. Use img2img for upscaling as you can use up to 8 steps and do multiple variations with different seeds. Try prompting low contrast and/or low saturation.
@SubtleShader Hell, even low steps img2img is not a good quality option.
@averyoncashius962 Then use 8 steps in img2img.
@SubtleShader Upscale is necessary, even for small monitor.
@SubtleShader I'll admit I don't know the difference between hires fix and img2img upscale under the hood. I was always under the impression that they would give similar results, but never tested it.
I'll give that a try and see how it works.
@averyoncashius962 @ChillDesire (I'm on Comfy) - 6 steps on the first gen. Upscale 1.5x with bicubic or a preferred upscale model. Img2Img with 4 steps and 0.5 denoise for lots of detail, toss out the ones that get deformed or reduce denoise if you really like the image. DMD2 looks incredible on some images and worthless garbage on others. Try a lot of different backgrounds, lighting, and outfits. Some are just going to look better.
@tsra In Forge 8 steps for txt2img and img2img are usually fine. Only sometimes I reduce to 6 steps in img2img. ComfyUI renders it a bit different.
I have also found Photo2 to be a little "burnt/harsh/sharp" compared to Photo1, so I've been sticking to 1 for now. (Not using DMD2, using DPM++ 3M SDE Karras @ 1440 x 960/810) @ 35 steps).
@SubtleShader can you make a simple img2img workflow like you made for your checkpoint? i love your simple 2 ksampler workflow you made but an img2img one would be great.
@JetterJack Sure. But you can easily do yourself. Just replace the txt2img stuff with a Load Image node. Will upload today.
@SubtleShader brother it's just I'm new and comfyUI is so complicated with these nodes, would appreciate your help
@SubtleShader i'm a noob lol. thank you. did u upload it?
@JetterJack I uploaded it here: https://civitai.com/images/106855958
Also added it to the Big Love description.
About the Photo2 version.
I tried in A1111 but to no avail. The results were.. disappointing. Sorry. 😥
As dissappointing as my Photo2 showcase images? As dissappoining as all other Big Love versions? Not sure what you expected... or if you used some strange settings or bad prompts.
To be fair to the creator, this model has amazing potential but needs a little "massaging" if you're coming from some other models.
Once you learn to tickle it just right (by following the instructions to a T), you get good quality images and a model that understands concepts many other models don't.
If your results are poor, I highly suggest you work with the settings a little more and double/triple check Subtle's instructions. They were vital for me to get good outputs.
What were you trying to achieve ?
For immediate "top-notch" numeric porm, use DMD lora 8 steps LCM+expo.
For more "arty" stuff, sde+karras 25 steps, then img2img upscaling 1.2 or 1.3 denoising 0.5/0.6 with or without DMD.
These BigLove merges are the less disappointing models around here ...
@SubtleShader In SDXL I mostly use Big Love XL v2, 2.5, 3, or 4, but my favorite is v.2. It's simply the best for what I do.
On Pony I use Big Love v.2 and 3. And I'm very happy with them.
I'm running the A1111 on an AMD Ryzen 9 5950X, with 64GB RAM and an RTX4090 GPU, under Windows 10.
Still on top... But I find very "disappointing" this "Insta" model still does NSFW stuff... Can't you find a way to eradicate all these filthy pussies from our lives ?
No pussies, no humanity. The fundamrntal facts of life. Still some people like to censor them and extinguish life on earth. Shape shifting alien invaders that control credit card companies?
@SubtleShader pussies have been created to control us, how can't you see that for God's sake???
Thought you were a clever guy... It is very "disappointing".
... And I don't even talk about their anus... The smelly burning lair of Belzebuth himself...
@BretChampagne I'm having fun "being controlled"
@SubtleShader I m tempted to say "get away from my holly life ! Germanic Nudist Satan !"...
... or "Germanic Naturist Satan"... Which is worse...
Can you make a checkpoint with no anus please ?
Don't like the smell...
Ok I promise I stop XD
LOVE YOUR WORK!!!! This is as good or better than Flux.1 dev. Can you tell me if I can make one of my LoRAs that I have on my profile work with this? Or should I just use this and train it on my dataset as a CheckPoint? Thank in advance
You only have Flux and Wan loras on your profile, so you need to retrain on SDXL. Best use Big Love as the base checkpoint when training. I recommend training with the SDXL lora preset of OneTrainer (https://github.com/Nerogar/OneTrainer).
@SubtleShader Thanks brother! Are you any good with KohyaSS? I have tried to train my LoRAs on Flux.1 dev Krea, but without luck.
@Duck007 Krea is just a Flux finetune so should be trainable like Flux. Try OneTrainer.
@SubtleShader Thanks man! I will give it a go
@SubtleShader I just installed OneTrainer after your advice. If I may ask, to train Flux.1 dev Krea LoRA, do I need to select LoRA or Fine Tune?
@Duck007 You wrote you wanted to do loras.
@SubtleShader Yes correct 👍🏼 Just was checking what “Fine Tune” means
@Duck007 Are you the teen on your profile pic?:)
@kennithjasher609 Bro
Less saturation
This lora reduces the saturation without reducing image quality: https://civitai.com/models/1363766?modelVersionId=1540766
Other method for Forge, A1111 users:
How can I get them not to look at the camera while writing a prompt, like it's a secret film, like it's being filmed without them knowing?
Looking away, looking down, looking to the side, closed eyes, head back, facing up, backside. Just to name a few.
How can I do Inpainting withthis model?? please help me, my brain is going to explode
There is a Inpaint option on the img2img tab in A1111/Forge. For ComfyUI you need a inpainting workflow. I prefer to use Photoshop, create a selection, paste it into Forge img2img, upscale it 2x, paste the best result back into Photoshop, downscale it to 50%, then paste it into the selection and use the brush tool on a layer mask to hide the transition.
I use A1111 with Soft Inpainting. With this method you can continue using the same Big Love model and DMD2 Lora without a specific inpainting model.
Set Mask Blur to 16 for large areas but reduce if you are modifying fine detail.
Tick Soft Inpainting and leave default settings.
LCM Sampling Method and 1 Cfg Scale.
@SubtleShader Thank you very much, I will try all the methods
Congrats for the 100k downloads!
Hi folks. Any suggestion for the best ComfyUI Inpainting work flow to use with this model?
These models are the best of the best and keep improving! However, I still end up having to use loras for certain things, e.g., extra pubic hair and large labia. Would be great to have these included properly in the base model. I have also found that the ANC loras often improve image quality, so might be worth it to include in the training material.
Thanks. A lot of pubic hair and large labia are not my things, but I plan to train on such details in a future version. Still trying to improve existing concepts. Never heard of ANC loras. Link?
@SubtleShader i would train a bit more "sagging breasts" concept into the model personally. All of these models (I dont mean yours specifically) always give fakeish looking breasts with small nipples etc. It's hard to generate diverse breasts overall. Like the nipples are always perfectly centered and the same proportion on similar shaped breasts. I find this on like every model in existence more or less. Also i couldn't get photo1 to do anal, like at all. Maybe thats different in photo2 or insta I've yet to try.
@TheNecr0mancer Yep, more diverse pussies and breasts is always good. Anal is rather easy with all Big Love versions. To quote the Photo1 article:
"The default hetero sex prompt is: 1girl, 1boy, anal/vaginal sex, penis.
Photo1 also knows these positions: missionary position, (reverse) cowgirl position, doggystyle, spooning, pronebone position, straddling and (reverse) suspended congress. Anal-only: bridge position, piledriver position. Vaginal-only: double vaginal. Double penetration was also trained."
If it does not work good enough, you can still add my anal sex lora. See https://civitai.com/models/1444070?modelVersionId=1919016
Fuck me. I need to cover pussies, to not see that fucking labia. A simple sfw generation gives 60 years old pornmommy vagina, its depressing.
Here you are, its all for you, not a single one for me
https://civitai.com/search/models?baseModel=SDXL%201.0&sortBy=models_v9&query=pussy
@SubtleShader Yeah, fair enough, but it's nice to get variety and diversity, and your models are already great with that but could be even better! I have used these ANC loras (especially flat and saggy breasts) a fair bit and with great results combining with your models: https://civitai.com/collections/7306121
@Foucault69 Thanks. I only merge loras at a very low ratio which does not make them work for such specific stuff. I can only train it to avoid a degredation of overall image quality of the checkpoint. Will surely do it in future. Not there yet. You can use loras till then.
Hi SubtlerShader, do you have any estimate of when a new version of the BigLove checkpoint will be released? Will it be after the early access for Insta1 ends?
Might take me a bit longer this time. Getting a bit too stressful releasing every 15 days. But I'm already training Photo3. Should be trained in 1-2 days. Then I need to do tests & maybe more training. Creating showcase images also takes 2 days at least.
@SubtleShader you're an absolute machine.
Absolutely brilliant! Thank you so much.