From 65dff3da4ba6c278fb6fa4b25d13de82c2805174 Mon Sep 17 00:00:00 2001 From: cmdr2 Date: Wed, 9 Nov 2022 15:37:51 +0530 Subject: [PATCH] Updated VAE Variational Auto Encoder (markdown) --- VAE-Variational-Auto-Encoder.md | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/VAE-Variational-Auto-Encoder.md b/VAE-Variational-Auto-Encoder.md index bd2fe49..afd9922 100644 --- a/VAE-Variational-Auto-Encoder.md +++ b/VAE-Variational-Auto-Encoder.md @@ -6,4 +6,9 @@ In short, it improves the generated images. A VAE is trained for certain aspects Stable Diffusion UI installs and starts using the `vae-ft-mse-840000-ema-pruned` VAE file that works with all Stable diffusion models. This uses the official one: https://huggingface.co/stabilityai/sd-vae-ft-mse-original -To use a custom VAE file, you can place any number of VAE files ending with `.ckpt` or `.vae.pt` inside the `models/vae` folder, and refresh your browser UI. After that, you can select the desired VAE file in Image Settings. \ No newline at end of file +## How to use a custom VAE? +Put your VAE file ending with `.ckpt` or `.vae.pt` inside the `models/vae` folder, like this: +![image](https://user-images.githubusercontent.com/844287/200801595-17276b3a-62e0-4d06-bfe1-a6481ba7921d.png) + +Then refresh your browser UI. After that, you can select the desired VAE file in Image Settings: +![image](https://user-images.githubusercontent.com/844287/200802028-d35ceda1-8816-4d5a-958e-168c6ad3fbd5.png)