Visit Civitai-Crystal team on discord! -news, updates, development, beta testing
Join RunDiffusion on discord! - Auto1111 and InvokeAI in the cloud.
------------------------------------------------------------------------------------------------------
In the attempt to break the barrier between what's digital and what's real, one model rose to the challenge:
Crystal-Clear-Photorealistic
Now you can generate images as realistic in appearance as your prompt allows it.
This model was not developed for something specific so feel free to try out different things.
From regular portraits to sci-fi aliens in the 5th dimension. It should be able to generate a full body view of things at 704x1088(raw) without any mutations or duplicates.
You can go far, with your creations, by having a good prompt that has a good negative to it as well.
This model is the accumulated result of intense testing, training, feedback and minute detail improvements, upon the Crystal framework, done by an entire team of people over a 3 week period.
__________________________________________________________________________________________________
If your computer is not strong enough, a solution to your problem is RunDiffusion where you can generate your favorite CIVITAI models or you can train new models/LORAs or embeddings.
The model is based on the Crystal-Framework which was created to produce clean, noiseless and unrivaled image quality, while adjusting the colors and sharpness of the image. Although, CC-Photorealistic cannot compete with Crystal-Clear when it comes to producing proper content at very high resolutions, it is unrivaled in the realism domain. Be it food, houses, interiors, animals, nature, photography the choice is yours.
Future versions of this model will tackle the resolution issue !
Note: To produce high realism images, it is advise to use a low CFG ( between 4 and 7).
The model was not tested for mature content or how well it produces it !
Feel free to post your images and feedback, also, don't forget to join the Discord (extra goodies and models available):
Visit Civitai Crystal-Clear on Discord!
If you do plan to include this model in your creations, please show us what you have accomplished! We will also help promote your model if you give us the deserved credit.
Description
Crystal-Clear-Photorealistic purpose is to create an image that is as close to realism as possible.
Based on the Crystal framework trained locally.
FAQ
Comments (13)
Awesome!, Will you be making any more Crystal clear type stuff, (maybe a anime/2d one)?
Hey! have you checked out his CrystalClear2 model? Another fantastic model from WarAnakin, which does anime quite well!
Well this gives some pretty dope results!
Fixed a misspelling error saying "crystalclear_v10" instead of "CrystalClear_Photorealistic"
Without having to go to some other platform, how about you list out the models used, so we can be informed of licensing, caveats, etc.
Or just, to do the thing so we can just generally have information.
You should definately add some prompts to these examples
i will bro
it should be mandatory to add prompts/models used in samples
Downloads as a LoRA rather than a Checkpoint file, am i missing something here?
Hi, that's a 1st. What's the filename it's saving as ?
crystalClear_Photorealistic.safetensors
@iamziyah89631 why do you say that's a lora? it's the name of the model
@WarAnakin I'm an idiot. for some reason, i thought checkpoint files had the file name ".ckp" not ".safesensors" i always corelated the LoRA files were exclusive to the safesensor file name. lol im still new to this
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.