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! You can get Klein2 for $10 on Tensor Art. A subscription includes the latest release and new ones within 30 days.
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 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.
The fp8 mixed version is for 24+ GB VRAM (highest quality), the nf4 version is for 12 and 16 GB VRAM (25% faster), the GGUF Q5_K_M version is for 16 GB VRAM (softest) and the nvfp4 version requires 24+ GB VRAM (on RTX 5090 2-3x faster, otherwise 2x slower, lower quality than fp8 mixed, less saturation, textured).
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.
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 Klein is a fine-tune of Flux 2 klein. It can do txt2img but also edit images with prompts. Txt2img is currently experimental and requires more prompt & seed hunting. Image editing works stable. It only needs 4 steps, so is quite fast (5 sec on 5090, 15 sec on 5070). You can run it locally with ComfyUI and online on Tensor Art. Use the GGUF Version with 8 GB VRAM, the fp8 version with 12-16 GB VRAM and the (pruned) bf16 version with 24+ GB VRAM. The nvfp4 version (full nf4) renders 2-4x faster on RTx 5xxx GPUs with the latest CUDA drivers. The base version (full fp32 or full bf16) is meant for training a lora only. 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. Forge Neo is not recommended, because it currently produces too low image quality with Klein, which will hopefully change soon.
Settings are 4 steps (2 steps for upscaling), cfg1, Euler sampler, Beta scheduler, 832x1216, 1024x1536, 1280x1920 or 1536x2240 pixels
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 (dated July 26, 2023), the Big Love Z-Image and Ernie versions are licensed under Apache 2.0, and the Big Love Klein versions are licensed under the FLUX Non-Commercial License (Black Forest Labs Inc.), 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, 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
Photo5 produces more photo-like images, is compatible with more types of prompts and offers 25 new concepts, e.g. various photo art styles, candid, pinup and cuteness concepts (See description!). It was trained twice on 2100 new images. The final version was selected among 17 merges.
FAQ
Comments (54)
Hello SubtlShader ! Will you cancel the download of FP32 Diffusers version of Big Love ZIT2 model after ZIT2 model being free ( after 5 days)?
You mean if I will remove the fp32 download? No. Why should I?
@SubtleShader OK, thank you so much
Can you generate pictures of a male with male characteristics?
Yes, of course. You can see men in some of my images. I also generated men-only images, but I usually don't post them, just like most people don't here.
@SubtleShader ok
i feel like 4.5 has better face diversity than 5, and the images are subjectively better on the same prompts
How many prompts did you test? I compared 300 different prompts between the two versions. My results say the opposite. Face diversity looks similar or higher than Photo4.5 to me. Besides, Photo5 knows a lot more concepts.
Some prompts always look better with a certain checkpoint. Why I test so many prompts & generate around 3000-5000 test images to compare. I've also developed a rating system to evaluate the changes. Results are quite clearly in favor of Photo5.
is there a way i can use the same enviroment , like the same bed and room ?
Use a prompt template for them, but don't expect 100% consistency. Additionally train a lora if you want more bedroom consistency.
Try training with some images created with the background you want; for example, add 3 or 4 images created with Photo5 and train again. The new model should give you the default background when you create images again.
What training resolution and image size is ideal for Photo5 character lora? 1024x1024?
1024 training resolution. Images can be 832x1216, 1024x1024 or 1216x832. Or similar sizes with other aspect ratio.
Question.
Z-Image has 8/10 ultra weird mutations.
As example on "A woman sucking a cock of a man" its 8/10 a man holding a cock in his hand and putting it in her mouth.
Our much worse stuff.
Nothing we can do against it i guess ?
I guess this happen through the distilled behavior, right ?
Will you finetune/train a lora on z-image-omni-base when its out ?
It has nothing to do with distilation. Z-Image-Turbo was just not trained enough on that stuff. ZT2 does it better, but still requires more training, which I intend to perform in future. I will finetune Z wether base comes out or not.
I also plan to finetune Flux klein 4B. Actually looks almost more promising than Z as it is smaller & faster and already has editing capabilities. A first training test looks good.
@SubtleShader but just to make sure. Its not my wrong comfy which produce mutations or gives bad prompt following ?
@LDWorksDV No idea what settings you use. But Z does porn with deformations. Big Love ZT2 is a lot better at it.
The drug cartels are coding all the NSFW models. Lost count of the number of images of guys with their cocks cut off, shoved down their own throats. 🤣
@InsidiousOne Okei yaa because i have the same problems xD
@SubtleShader I used your workflow of z-image and got those deformations.
@SubtleShader ZIT does do a better job with some scenarios but the one thing it lacks is imagination. It'll follow a prompt, but it'll give the same results with every seed. It's like a dog with a bone. Your model is solid, it's just ZITs lack of imagination.
@InsidiousOne thats dream scenario, if you use wildcards for everything tip to toe.
@InsidiousOne Big Love ZT does different images with the same prompt. Unlike ZiT.
@LDWorksDV Sorry to hear. I mainly use Forge Neo and only get deformations when using a batch size higher than 1. So I stick to batch size 1 and have no problems.
@emirlennard772 I haven't graduated to a webUI yet, I'm using Stable Matrix's Comfy app locally. I haven't figured out if wildcards can be set yet.
@SubtleShader I have noticed that your model does in fact do very well at changing faces with each seed but the composition doesn't change much. Have you considered training a Chroma version? It seems to be the best at sticking to prompts while filling in the blanks with semi-random composition that makes sense. I just have trouble making the images appear photorealistic with Chroma and that it's sloooooooow on my 3090.
@InsidiousOne i dont use comfy, but according to LLM, search in Comfymanager for DynamicPrompts node.
@InsidiousOne I'm not a fan of Chroma. Will train Flux klein & Z next. Flux klein is very promising.
@emirlennard772 I'll look for the node you mentioned and see if it shows up in Stability Matrix after install. Thanks.
just have trouble making the images appear photorealistic with Chroma and that it's sloooooooow on my 3090.
Chroma seems to be heavily finetuned towards anime (like Pony/Illustrious) and falls into it when it finds danbooru tags. I can recommend you this LORA to prevent fall to anime or use UncannyPhotorealism checkpoint or GonzaLomo Chroma checkpoint.
@mphobbit Thanks. I don't think sliders work with Stability Matrix but when I switch over to the webUI I'll try it out. I have both the other checkpoints and only dabbled with them. Today I tried the Flash Heun r64 w/V48 Detail Calibrated and it sped things up A LOT. Also testing out Lenovo and Prof Photo loras and they both kick realism up another notch. An AI search also allowed me to cut and past specific cameras/lenses and also suggested not using terms like "photorealistic" etc. Chroma likes natural speech and not tags. Still trying to find the best sampler/sched combo but I'd say all this brought the photo quality to at least Pony level. I made good progress today with everyone's help. Thanks!
@InsidiousOne
> Chroma likes natural speech and not tags.
It's a popular thought but personally I disagree with it. This image was made with NLP but Chroma's general dislike sci fi topics made picture more artistic. And this one was made with tags.
According to my experience, Chroma falls the image towards anime/semi in case if you add something anime or cartoon-related overriding any positives and negatives related to photorealism. For example: this picture doesn't contain a cartoon hero, and this one contains. Both are made with tagging.
Still trying to find the best sampler/sched combo
Try Euler/Beta and DPM2/Beta with CFG 4.5-5.
@mphobbit Yes, I think Chroma leans to cartoon/paintings too. The reason I'm not giving up on realism hopes is that it "seems" easier to fine-tune prompts whereas other bases can be just so stubborn sometimes. The Prof Photo and Lenovo Loras definitely help with realism in Chroma.
My last batch was Euler/Normal at 2.25. The Euler/Beta is dimmer and loses a little detail, I'll play with it. DPM2/Beta loses half the prompt adherence at 2.25, but if I bump it up to 3 it starts coming back. But at 4 it's way too much contrast and cartoony. I'll keep at it. Thanks for the tips.
Is the fp32 version a .zip file? After decompression, there are several files that cannot be loaded and used on comfyUI.
Yes, it is diffusers format, so multiple files in a zip. It is not meant for Comfy but for training on it. Use the bf16 version for Comfy.
Anyone trained a character LORA with this checkpoint yet? How did it turn out?
You should specifily if you mean the ZT, Photo or Ultra version. It did character loras on Photo & Ultra so far. Were great. But also seen people do bad character loras by using training images with low quality. A guy I know does character loras on Big Love Photo for clients.
The character lora only needs good portrait images. Does not need to be NSFW. Afterwards if you use the lora with Big Love, it will produce any type of image with the character.
@SubtleShader Oops sorry my bad. I meant ZT2/ZT1
@SubtleShader i fokkin love that! BIG FAN LOVE! Say i make a few character loras of famous folks. Its all good as long as i dont publish it or gain something correct? Thank you
@Samartdelete As long as you do not publish the lora or the images that were produced with it, you can have your own private fun.
Future Projects
Now that Ultra improved quite a lot, it seems to be more popular than Photo early-access-wise, so I will focus less on Photo in future. Ultra4 coming soon.
Instead I will focus on Flux klein 4B which offers a open-source license and is 2.25x faster than klein 9B and 50% faster than Z-Turbo. With its speed, higher resolution & editing capabilities it actually looks more promising than Z-Turbo. Only obstackle is its undertraining, which can be resolved.
I plan to do a full finetune on both klein 4B and Z-Turbo (even if no base released) with the same dataset.
Getting only a grey image on Forge Neo :(
Check your settings. Sounds like you don't have the DMD2 lora installed or you do not use it.
I usually change the checkpoint model, render one and swap back and it usually works. Or take out prompts that you added if it was doing fine before. I get the gray images occasionally
First off all thanks a lot for your great effort.
2 part question, I have a hard time getting good quality with both 8step, dmd and full 25 step often the faces feel out of focus or in general the whole image is just a little off (ive used recommended settings).
Secondly its even worse with my own character lora (trained as recommended and with my own training recipe that proven good results on 10x other checkpoints.) but on photo 4.5(and xl3 and others from here) i either get blurry faces, overall low quality and softness.
any suggestions would be much appreciated!
Can you post a image with metadata and give me the link?
Thanks for your quick reply.
After much trail and error i think i ended up working all the dials correct. leaving comment for anyone else that would need help. maybe obvious, im not really an expert but:
lower your DMD lora strengths -> creates softness but better likeness which means you can push your lora (tags:1.X) and strengths less with same face likeness. use detailer or other second sampler up-scaling methods to make it more natural, realistic and less blurry
First of all, thank you for your contributions; what I've seen is fantastic. I'd like to ask a question out of my ignorance: among the Ultra, Photo, and ZT versions, which one will give me the most realistic results? I'm looking for realism in skin and genitals. Thank you.
I think you will want to wait just a bit, Z-Image (base) was released 5 hours ago, so we are about to find out what the non-Turbo model can do as far as fine-tuning
Ultra is the best of the 3 in terms of realism. Main advantage over Photo is the higher resolution which makes images look more realistic and detailed. Photo on the other hand is still more mature as far as concepts & composition is concerned. ZT is less realistic but is able to generate complex things and is more prompt adherent. ZT has least realism concerning genitals.
@SubtleShader Thank you so much for take the time to answer my question! Ultra versions will be my goal then! THANK YOU!
@ralphjx Ultra has less stable anatomy than Photo, so you need to generate more. But always depends on the prompt. Some prompts produce very stable results.
Z Base is out! Can’t wait to see what you’re cooking, shader 😏
First Flux klein 4B (image editing capabilities!), then Z base.