Categories: Tutorials

Enhanced Resolution: 8x Upscaling

When it comes to upscaling the image in Automatic1111, and if you have low VRAM you must struggle a lot it doing this. Now, they are a methods that will be going to solve your problem. We are using this extension and is quite effective for upscaling. 

There are multiple ways to upscale like hires fix, and ControlNet tile. It just upscales the image without adding extra detailing. But, it’s a Multidiffusion extension for Automatic 1111.

Features included:

  • Stable Diffusion XL model supported
  • StableSR support
  • Tiled Noise Inversion
  • Tiled VAE
  • Regional Prompt Controlling
  • Img-to-Img upscaler
  • Ultra-scaled image generation

Let’s see how to work with this and convert our art to the next level. But, first of all, let’s dive into the installation part.

Installation:

There are basically two methods to install this extension. the process will be as simple as you install the other extension.

(A) Method 1: Install from Automatic1111’s download list 

1. You need to get your Automatic1111 installed to use this extension. 

2. Start your Automatic1111 WebUI and go to the “Extensions” tab then select the “Available” tab then click “Load from” and search for the “[TiledDiffusion with Tiled VAE]” extension.

3. And at last click “Install“. It will take some time to download the prerequisites.

After installation restart your Autoamtic1111 by clicking on “Apply and restart UI” to take effect.

(B) Method 2: Install from the official Github repository.

1. Go to your Automatic1111 WebUI and Select the “Extensions” tab. Now move to the “Install from URL” section and copy and paste the Git URL which we have provided below and at last click Install :

https://github.com/pkuliyi2015/multidiffusion-upscaler-for-automatic1111.git 

2. After installation restart your Autoamtic1111 by clicking on “Apply and restart UI” to take effect.

How to Use this extension?

For using this extension, we don’t need to use extra settings. Here, we have to use the “Image to Image” option.

1. Generate your image by inputting your positive and negative prompts. Here, we have generated an image of 512 pixels and we will upscale the image to 4x.

2. Now, click on the button to send the image to the image2image tab available on the right-sided canvas of Automatic1111.

3. Here, in the prompt box remove the old prompt with upscaling prompts. Like highly detailed, uhd, 2k, raw photography, etc.

2. Set up your settings as usual for image generation. We have set it up as follows:
Sampling method: “Euler
CFG scale:1
Seed: -1
Denoising strength: 0.2
Here, do not try to change the image dimensions.
3. Into Tiled Diffusion extension tab:
Enable the Tiled diffusion option. Then select the method “MultiDiffusion” and Upscaler as per your requirement we have set it up to 4x. And enable the “Noise Inversion“.
Again turn on the “Tiled VAE” option. Set up your encoder file size. And at last hit “Generate“.

Now, you can see we have upscaled the image to 4x from 512 pixels to 2048 pixels. You can see the added detailing is very impressive and incredible.
admage

Share
Published by
admage

Recent Posts

InstantIR: Restore Your Images

Figuring out the model that can fix your low quality pictures? Now, restoring your low…

4 days ago

Omnigen: Next-Gen Image Generation & Editing Tool

Traditional diffusion models uses various mechanisms for image modification like ControlNet, IP-Adapter, Inpainting, Face detection,…

1 week ago

Create Engaging Videos with Mochi1

Mochi 1, an open-source text-to-video diffusion model has been released by Genmo.  Trained with 10…

3 weeks ago

Local Installation of Stable Diffusion 3.5

So, it's finally here. Stable Diffusion 3.5 has been released by StabilityAI on October 22nd,…

4 weeks ago

Video Depth Mapper: Efficient 3D Depth Mapping Solution

Due to the extreme diversity of video content like motion, camera panning, and length, the…

1 month ago

Top 30 Negative Prompts for Reliable Diffusion Models

 You can use these negative prompts while generating images using Stable Diffusion models.1. Clear BrandingPrompt:text,…

1 month ago