Try out the Noise Offset version of the 2.0 and 3.5! Here
If you like my work, consider buying me a coffee ☕
V4.1 (experimental):
removed rev_animated
added realistic vision
fixed the model
Recommendation: clip skip 1 (clip skip 2 sometimes generate weird images)
2:3 aspect ratio (512x768 / 768x512) or 1:1 (512x512)
DPM++ 2M CFG 5-7
prompts that i always add: award winning photography, Bokeh, Depth of Field, HDR, bloom, Chromatic Aberration ,Photorealistic,extremely detailed, trending on artstation, trending on CGsociety, Intricate, High Detail, dramatic, art by midjourney, masterpiece, best quality, high quality,extremely detailed CG unity 8k wallpaper
V4.0:
added deliberate 2.0
added rev animated
fixed hands with basil_mix
V3.5: ADDED Pastel mix
https://civarchive.com/models/5414/pastel-mix-stylized-anime-model
V3: ADDED AnythingV3 and fixed the model: https://civarchive.com/models/66/anything-v3
RECOMMENDED VAE'S
kl-f8-anime2
Anything-V3.0
RECOMMENDED LORA'S FOR BETTER RESULTS
0.8 weight
V2: ADDED Deliberate: https://civarchive.com/models/4823/deliberate
Roboetics: https://civarchive.com/models/3738/roboetics-mix
Inkpunk: https://civarchive.com/models/1087/inkpunk-diffusion
DreamShaper: https://civarchive.com/models/4384/dreamshaper
ChromaV5: https://huggingface.co/SomaMythos/ChromaV5
H&A's 3DKX 1.1: https://civarchive.com/models/2504/handas-3dkx-11
Analog diffusion: https://civarchive.com/models/1265/analog-diffusion
The Ally's Mix: https://civarchive.com/models/1202/the-allys-mix
Tokens: Analog style, chromav5, nvinkpunk
Chromav5 Keywords: Chromatic Aberration; Geometric Shapes; Bokeh; Depth of Field; Photorealistic; Cosmic; Detailed; Bloom; HDR
Description
FAQ
Comments (10)
I don't think it was a good idea to merge rev_animated it's heavily weighted towards females. V35 is still better imo. In v40 it's a lot more difficult to keep females out of the images now. About once every 5 -10 images I get a big female face appear, even with negative weighting.
yes, i see that, now im merging without rev, and adding realistic vision, probably in a few hours ill upload a new version, thanks for the comment
Could you give me an example of the prompt you used?
How does V41 compare to V35? I'm wondering whether I should update my model on file
IS 2048 Hires?
Any vae to use along with this checkpoint? Also, any comments for compatibility with LoRA like LatentLabs360 and 360Diffusion?
Despite the fact that this realistic model is already very old and has some instability, problems with anatomy and composition, it still produces quite good original pictures.
I'm not too surprised at a landscapes-focused checkpoint being bad at anatomy.
Does this version use any special VAE?
If it doesn't already have one baked in, grab anime pastel vae or clear vae from any other sd1.5 checkpoints that can be downloaded on civitai that have the vae as a separate, downloadable file, those two are good enough. If you end up wanting less saturation or more, you can always take care of that using Krita or Gimp.