For business inquires, commercial licensing, consultation and large scale image captioning for datasets contact me under [email protected], for custom LoRAs visit my ko-fi page.
Checkpoint Type: SDXL, Realism and Realistic
Follow me on Twitter: @YamerOfficial Discord: yamer_ai
If you have enjoyed this checkpoint:
Yamer's Realistic is a model focused on realism and good quality, this model is not photorealistic nor it tries to be one, the main focus of this model is to be able to create realistic enough images, the best use with this checkpoint is with full body images, close-ups, realistic images and beautiful sceneries, for example futuristic cities or insides of a home.
Give this model a chance and let your creativity guide you and the perception you have on the world!
This model is not nsfw focused but it can crate nsfw content, it is preferrable to use a LoRA for specific nsfw content, I also recommended to use words such as: Realistic, Ultra realistic, Photorealistic etc..., to give the resulting image an even better and more accurate realistic look to it.
If you want to explore fields that are not realistic focused I highly suggest to use my LoRA for abstract work and a better ethereal look: Yamer's Style.
No refiner is needed for this model to produce good quality results.
Adetailer is not required to achieve a realistic looking image.
The main style of this model is realistic but that can change with a specific prompt that aims at an anime look for example.𝐕𝐞𝐫𝐬𝐢𝐨𝐧 𝟓 𝐮𝐩𝐝𝐚𝐭𝐞 𝐧𝐨𝐭𝐞
After some time I decided to make some change in the way I release models, the new version will be released in 3 separate models, these 3 models while being similar can still be better with some prompts, that is why it is up to you to decide which is better.
The new version 5 is separated in: TX, SX and RX.
One version can be better with colors or saturation than the other 2; The prompt used in this example is the same, steps, seed and sampler were the same for all images.
Version RX (Realistic + RunDiffusion), SX and TX have been uploaded, enjoy all the current versions of the checkpoint!
Special thanks go to the RunDiffusion team that helped with the release of Realistic 5 + RunDiffusion Photo! Show them some love :3
Version 1, 2 and 3 have the SDXL VAE already baked in, "Version 4 no VAE" does not contain a VAE; Version 4 + VAE comes with the SDXL 1.0 VAE already baked in.
V5 TX, SX and RX come with the VAE already baked in.
Thank you for using my models and writing a review, all forms of support are appreciated, it takes me a lot of time to produce this kind of checkpoint every time! ❤️
Recommended settings:
Image resolution: 1024x1024 (standard SDXL 1.0 base resolution)
Steps: 30 to 150 (Lower steps may produce image artifacts and other unwanted results)
Restore Face: Not needed to generate high quality images (Not used in the showcase)
Clip skip: 1 and 2 (Works good with both settings)
CFG: 5 to 20
Hires Upscaler: 4x_foolhardy_Remacri or 4xUltraSharp
P͓̽e͓̽r͓̽s͓̽o͓̽n͓̽a͓̽l͓̽ ͓̽R͓̽a͓̽t͓̽i͓̽n͓̽g͓̽:͓̽
Quality: 9/10
Realism: 8/10
CFG resistance: 15/30 (going above 15 may result in bad images)
Hires steps: 10-50
Style: 8/10
NSFW: 7/10
1) If you look too closely at the form, you'll miss the essence.
2) Art is not perfect and will never be, it can only improve overtime.
3) I don’t paint dreams or nightmares, I paint my own reality.
-Tally Hall -Yamer -Frida KahloDescription
Thank you to RunDiffusion for giving me the opportunity to partnership with them!
The model has been merged using Realistic 5 + RunDiffusion photo checkpoint.
FAQ
Comments (31)
I am creating since Days hundrets of the best expressive Pictures with this Model.
I thank you so damn much.
Prompts for Masterpieces (PERFECT Pictures) of female Angels with amazing effects... o.O
(+) Positive Prompt ► 391 Words,
(-) Negative Prompt ► 31 Words
Size = 1520x1280
Hires.Fix ► (4x Ultra Sharp)
Hires.Fix Denoise ► 0.60
Hires.Fix UpScale ► x1.35 -> 2052x1728
Refiner ► SDXL_v10RefinerVAEFix
Refiner Switch ► 0.69
CFG ► 14,5
Steps ► 36
Sampler ► DPM 3M SDE Exponential
PS: At the Beginning.. I had long time Troubles with wrong anatomy (Limbs, Fingers and Feet) After a lot of Generating.. I could turn it into a perfect stunning Picture Generator. So damn.. It is great.
KG RoF OV EVO
I'm really happy to hear that, I hope you like my other checkpoints too! :)
If you need help with prompting feel free to message me on discord: yamer_ai
@Yamer I will test everything for getting exactly what I imagine to. For Sure, I can and will help if needed. Till now, I wrote more than 148 000 Words in English for Prompting the Best Pictures of my Imagination. For Now, I think I generated over the last 6 Months more than 492 000 Pictures xD
You said there's SX, TX and RX but where's RX hiding?
I deleted my bad review. It should show up. I apologize. I suppose I'm just upset there are no good models for that type of thing compared to 1.5. Perhaps, if I make you mad enough with a bad review I can influence the direction of these SX models xD. Please consider making a model you're not responsible for, if you know what I mean ;P. This is in reference to the 1 star review. Undeserving. It should revert soon if it hasn't already.
Thank you for taking the time in reconsidering the review, I really appreciate it 😸
This model really looks great, the best realistic model soo far
😸
Couldn't agree more
@RunDiffusion ;)
Excuse my ignorance, but what is the "Rundiffusion" part of the checkpoint for? I did an article search but can't find anything that would indicate what one would use a "rundiffusion" for. Are you just noting it was trained on their website?
This model was a 70/30 merge with the RunDiffusion Photo model (a very realistic checkpoint), realistic 5 was used as base being the 70% and RunDiffusion photo being the 30% making this model even more realistic, I hope this clarifies how this model was made.
Hello,
I just came across your model on mage space and thought I'd try running it locally.
Could you please help me out with the settings i should use to achieve output as on mage space please. I see you have mentioned somethings but not all such as which DPM etc. to use for best resuts.
My query being for the Realistic 5.0 + RunDiffusion version
Thanks for your time : )
You can see the settings I use by looking at the metadata info of the images using png info on a1111 or by clicking the i icon on the bottom right part of all images in the showcase, if you don't know how to use png info please let me know, I hope this can help you achieve your desired result ;)
@Yamer Thank you very much for the quick reply. I understand how it works here now.
I did have some follow up questions please,
- I see the step and cfg count is quite high, and thus it takes a while to render a single image in my case(4080), this is unusually higher than some other models i have seen, this is simply normal for this model then? or can I go lower
- This is perhaps more of a general question, I noticed in everything i have tried so far getting a third or fourth person to appear in a scene seems to almost never work. Generally speaking is this a limitation of sdxl as such? or am i simply prompting incorrectly
I get this message when I try to use this model on Tensor- Generate failed model file sdxlYamersRealistic5_v5Rundiffusion.safetensors is deploying. Can you please check it out and see if you have the same problem as I do.
Hello,
I'm having some trouble using this model with Instant ID. Thusfar I have used instant ID with its recommend lower than usual 4 Steps and 2 CFG for 1.0 models and it has worked fine with other models even though they recommend higher for their regular usage.
If I try using Realistic 5 with Instant IDs recommended Sampling steps to the same 4 and CFG Scale to 2 it renders unusable images.
Could you please help me out on how I can use Instant ID with Yamer's Realistic 5 please : )
Thank you for your time.
I'm not really super familiar with instant ID, try using more steps and a higher CFG, between 4 and 6
@Yamer sure I'll play around some more ty.
and Instant ID being that thing where one can use a face and render it with prompts of choice and the persons face is applied each time. it's quite a lot of fun to take your selfies and travel the world : P
Great model. I love it! Runs fast on CivitAi on site generator and works good with most LoRa's. Thank you Yamer for creating this awesome model❤
I'm really happy to hear that :)
Great model with phänomenal Output 👍
Thanks @Yamer for creating and making available to us. Only after applying yours checkpoints for my prompts, I did get this cool scenery. Here my 10 cents to show my gratitude.
All of your models are great, is there any chance you will be releasing a Flux base model some day.
this one is dang near perfect
Great model, works well for portraits!
Best SDXL available. Almost foolproof. Spits really impressive images. Like all others, it struggle with secondary subjects' faces in the background if any.
"If you look too closely at the form, you'll miss the essence." belongs to Rumi (Yalāl ad-Dīn Muhammad Rūmī), 13th-century Persian mystical poet and Sufi master. "Tally Hall" is an US rock band and the quote is wrongly attributed to them.
Thanks for your work, about to test this checkpoint!
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.