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
qwen_3_06b_base.safetensors goes in ComfyUI/models/text_encoders
qwen_image_vae.safetensors goes in ComfyUI/models/vae (this is the Qwen-Image VAE, you might already have it)
FAQ
Comments (20)
Bring back Experimental Real Dream Pony -Flux Dataset V.5 Lora
I highly doubt that will ever happen (not even in a "classic" variant). You can get the model from civitai archive or use one of the popular alternative generation websites to use it. In the archive the author states why he left SDXL, Pony and never did an Illustrious checkpoint.
@Raininosi what was the reason?
@dow185 The newer Models are superior over SDXL from a technical perspective, but the ecosystem will need to grow which is where this model is investigating in.
@Raininosi I wish he'd give Chroma a try :P
Any chance BF16 can be added to the releases? Would love to see what it can do!
y second this
@amnesiacyc344 Third
Can this checkpoint be used with a character LoRA without changing its identity?
@sinatra Could You explain how such tuning is done? Or point into right direction?
Asking because I was trying, via creating Lora, to make Anima output more realistic results but failed miserably, nothing short to what this is capable.
Good checkpoint, but tbh does it make anysense to use Anima as a base model in order to create realistic style stuff?
I would bet they are trying to get the diversity of prompt adherence from anime with realistic.
Have you ever tried Illustrious realistic models ?
As mentioned above, Anima has great character understanding and can be prompted with both tags and natural language. It also picks up styles well, and I've gotten some decent results with realistic LoRAs so far. It's not going to match something like FLUX or Z-Image but being able to throw a few tags instead of a wall of professional photography terms is nice. Even Anima's creator has trained (non-public) realism LoRAs.
I guess it does for someone with very, very limited hardware. If I was trying to go for realism in low VRAM I'd go for ZIT fp8 or nvfp4 models but... I don't know, maybe style mixing in anima?
For me it's about prompt adherence, as others mentioned, it does it very well, IMHO better than ZIT/ZIB/H/F/CH and, depending on merge, as good as iLL/Pony/SDXL.
Its small and smart model. Thats rare. All smarter models are usually big and slow, even Zimage aint that small or fast or smart, its always "pick one".
If someone will manage to make fully realistic Anima, while keeping rest of it functional, then its pretty much "gold".
Anima can actually already do Photos Pretty well.
try (real life photography:5)
@mac2492 I have little experience with Anima, but generally I can generate similar looking realistic images, but overall miss a large amount of flexibility that Illustrious offers. I found that when I do subtle changes to a prompt the image usually does not change at all, f.e. I say "dark skin" or "latina skin", there is a marginal amount of brown in the result. If I increase the weight, nothing happend at all. I added a skin Lora which was either very brown or light. Illustrious has a very good Skin Color Lora that I use for nearly every image.
@Raininosi The problem that I do encounter is prompt adherence with custom models. Most of them are over-fit and cause issue as You did described.
Usually when I download new model I put it into compare workflow with set of prompts and look for prompt adherence. I look at the image and description and see if what I prompted is on the image. Most of the models are hit and miss.
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.