Real Dream Website (Faster download)
Available at TensorArt: https://tensor.art/models/899250342552672189/
Comfyui Workflows:
Anima text2image with Loras Workflow
Anima img2img (denoise) with Loras Workflow
Klein 9B Edit with Loras Workflow
Klein 9B img2img (denoise) with Loras Workflow
Z-image-turbo img2img workflow
Z-image text2img with loras workflow
Z-image De-Turbo with loras workflow
This model is capable of generating very realistic images that can confuse people's critical sense, please do not use it irresponsibly, harming ordinary people or public figures. Also avoid its use in electoral campaigns, promoting or damaging the image of candidates. Although the field of artificial intelligence is poorly regulated, its improper use can cause legal penalties even in countries with little regulation in the area. Be very careful to not commit crimes in your country and not to promote injustices against innocent people. I do not authorize the use of this model for Lora training using the face of real people without authorization from the biological owner. Do not create images that can embarrass and damage the public image of people. This can cause serious psychological harm, including deep depression. I do not authorize the use of images generated by this model for scam applications and the dissemination of false information.
During the fine-tuning of AI models, I only use synthetic images that do not infringe any copyright, personal image rights, or laws. The models are merges of open models from different sources and may contain diluted data remnants in the neural network. I am not responsible for possible data remnants since they are diluted, and it is not humanly possible to track the existence and origin of this diluted data.
I do not allow the upload of any of my models on Tensor Art. I only do not allow it there because the policy of duplicate models on Tensor Art prevents me from uploading again if someone has already done it.
I have enabled to automatically hide all NSFW posts for several reasons that are not worth listing. I hope you have patience and understand.
Description
This is a model with additional training that is not as stable as the standard model. It has different characteristics in image generation.
FAQ
Comments (26)
Love freckles on this model
can you convert this model to SVDQuant-int4
I waiting for getting better and better
This is the best model to-date, it does the best in getting rid of plastic skin naturally too. If you render enough images, I've been able to get rid of plastic skin. The skin looks really good on here and freckles are amazing. If you need realistic photos use this one with the latest FLUX version.
This one model SDXL-Pony16 is the best to this day as the most realistic. Shadows,skin details, lighting. But sometimes it doesn’t respond well to prompts. Hope it will improve in the future.
I do think it lost some of the Pony prompt adherence for its gains in realism. Sometimes I just use 15 and then go over it with XL 6 at .5-.55 CFG for realism finish.
The Obsidian V1 shows real promise. A bit hard to work with but seems very realistic.
We need a new approach for XL(pony, illustrious).
I have enough, when i try a good, almost photorealistic model and it doesnt know anything about booru tags, sex pose tags etc.
A guy tried to implement t5 to xl, but there is nothing to show, its worse, than clips.
When pony7 will release, we will have 64gb vram 200 bucks cards, so it will be much easier.
I paid 150 € for an 8 GB card in 2017 and I can't find an 8 GB card under 180 € in 2025, so 64 GB for 200 sounds like a dream.
half the time 17 turns literally anyone into a white woman... even furniture, clothes, or plants lol
So many faces look the same with the same expression. They are clear but they look like the samples. Almost all of them.
There was a bit of overfitting in this version, I'm already aware of it and intend to correct it
Interesting model. My sunny beach background became flashlight bedroom.
There was a bit of overfitting in this version, I'm already aware of it and intend to correct it
Is VAE baked-in on SDXL-Pony v17 and Pony Obsidian v1 versions?
Everything comes out too dark and it's really bad about locations. In a field seems to mean in a bedroom.
I have not observed darkness, but I use Lora that control these quite well. But Indoor happens more often (and is broken more often as well, like simple 3D from stairs)). Town keyword is ignored (as usual, need to be more specific like school / shops / market). Avoid to generic keywords like field, use grass field for outdoors. Also "outdoors" works. Indoors on the negative prompt stopped working. Generally negative prompt is broken, requires Lora to work (on the current Pony , sorry, I just saw that).
New Patreon available, if you were a member of the old Patreon page, I recommend signing up for the new one. The old one is disabled, I'm trying to resolve it with Patreon support but I have no hope of recovering it any time soon.
Thanks for the great model — I've really been enjoying experimenting with it. Excellent work and much appreciated!
This is a great model! Thank you
This model is somewhat challenging to work with. Realistic, but difficult to achieve the desired results. Images come out a bit darker than I had anticipated. It seems best used for specific use cases, not casual image generation.
For my uses your Pony 16 and XL 6 are far and a way my favorite checkpoints. I use plenty of Loras for my own needs and these checkpoints really help on some not perfectly trained models I made. Thank you!! Best results I've had in two years of trying dozens of checkpoints.