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
FAQ
Comments (36)
9/10
Extraordinary model!! easy to manage with simple prompt result glorious realistic image
Wow, great job with the Pony V3 revision! You managed to take an already excellent model and make it even better.
Real Dream now seems to be much better at expressing Asian women; previously it was pretty dead-set on a tanner skin tone (and adding tanlines without prompting). This no longer appears to be the case.
Also, it seems to generate much more Asian-looking faces when prompted for (and better faces in general) and produces less freckly skin by default.
Awesome work! This is now easily my favorite Pony realism model :)
I think this is the best of the realism Pony models. It's the best at: skin, hair, explicit areas of the female body, prompt adherence, to name a few.
https://civitai.com/posts/2906442
However, I'm having trouble producing other ethnicities/nationalists. I can get Asian or White. However, I can't dial that in to; Korean, Japanese, Nigerian, Black, Indian. Is there something I'm missing?
by far the best realistic pony model!
I was saving up my buzz for a profile background, but I thought if you could make Real Dream even better, then it's probably better given to you.
What's different in v4?
please update lcm!!! :)
You still need more buzz for training?
I'm accepting, with more buzz I can train with more steps than the standard. There is also the issue that a Lora can go bad for a number of reasons, so I need to train more than once. CivitAi has been the best platform for training SD, I just think it's bad that it doesn't support Lycoris.
Buzzed! Keep up the good work.
Please let us use this in the Civitai generator! 🙏🏻
Unfortunately I don't have the option to select which model and which version can be used in the generator. The people of Civitai decide and apparently it's an algorithm.
@sinatra Well that sucks. It worked for a bit and it was wonderful. Thanks for your excellent work!
What's the upgrade from Pony v4 to v5?
I did two subtraction merges that improve photorealism and increase the contrast of the images
What do you think of the current state of SD 1.5? Is there still a way to fine tune the model or have we reached its limit?
limit, the number of pixels mean things like fingers will never be "big" enough to get right at small scales. I do wonder why hires fix can't get that down better but it is what it is, Pony is the current leader for naughty things...
This is absolutely the bets realistic model I have used for what I need it for. I can't remember what version I'm running, is there a way to find out? I've tried SDXL but I must be doing something wrong when I'm making LORA because they never look even close to real as my lora with your model here. SD 1.5
It's done. It can't do human anatomy well no matter how well you train it. I've pretty much stopped using SD1.5 altogether now.
Didn't really tested the others, but just tested V5 pony a bit, seems it kinda lost the wild (mostly NSFW, I mean hardcore) element of the pony models, sure it looks more realistic, some even close to photorealism, but at a cost that IMHO don't worth the pony base, if you get a pony base it should mostly maintain the pony capacity, which don't seems to be the case here, So I would say not a grat 'realistic' pony model because of this, Zonkey 4 or ponyrealism v20 for example completly outclass this one
I didn't change the model that much considering the percentage that was changed, the focus was on improving the photorealism effect and the stability of the image. If you are no longer able to get the same results as pony, try changing the prompt writing approach because the data from the pony model is still in the model.
@sinatra Sure i'll give it some more test, but so far it looks like it doesn't really have the explicit NSFW stuff,
The NFSW is still there, but one have to force it using (hand holding:2.2) for example.
Any chance of pony version avaliable at TensorArt?
Pony is not working well on TensorArt
No fights guys, let's only mistreat the GPUs
Do you want a Hyper-SDXL version? React to this comment so I can know the demand
please give an updated LCM version!, although please don't make it so we have to have an LCM sampler
Yeah, Hyper all the way!
nope, I prefer normal sdxl instead of LCM or lightning, LCM / lightning are not very realistic and the outcome is buggy most of the time.
please give an updated lcm version! although please don't make it so we have to have an LCM sampler
What's the problem with the LCM sampler? Have you tried updating the Web UI? it has been officially added