metadata
author: Quiescence
baseModel: SD 1.5
hashes:
AutoV1: D6BD7129
AutoV2: F768813EDC
BLAKE3: 96BE1502EAA0A73DE02DEF1983D474499662281859B70B7582FD200CC8FDE731
CRC32: C57B09F9
SHA256: F768813EDCE84239D70CE8A24CB496B9EF444C4E5AAFF435E45CAAD5CC1D39A2
metadata:
format: PickleTensor
modelPage: https://civitai.com/models/296442?modelVersionId=333102
preview:
- >-
https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/cf5360ee-077e-4c1d-bc52-6b9f755a7572/width=450/6366795.jpeg
website: Civitai
Trigger Words
No trigger words
About this version
No description about this version
OrangemixVAE !reupload
this is a reupload of the OrangemixVAE.
The original model seems to have been deleted.
VAE-explaination
A VAE (Varitional Auto-Encoder) is a type of model that can convert latent "images" into pixel images and vice versa.
Using a seperate VAE over a VAE baked into a model can help with Oversaturation or Washed out images.
How2use
Automatic
Go to settings ->
User interface ->
User interface ->
Quicksettings list
search for: sd_vae and hit enter
put this model in
models\VAE
ComfyUI
put in models\vae
refresh if you have comfyUI open
loaders -> vae loader
select:
orangemix.vae.pt
connect to VAE en/decode nodes if neccersary.