Introduction

This guide covers how to upscale an image or bulk images-- whether it be Stable Diffusion output, comics, photos, etc.

Prerequisites

• This guide assumes that you have chaiNNer installed-- if not, you can follow the guide here.
• This guide assumes that you have a model. If not, you can grab this one for the purposes of the guide.

General Instructions

• We will be utilizing the pre-made templates which fellow community member Kim created. First, visit the templates page.
• The repo is organized into NCNN, ONNX and Pytorch. If you are using a NVIDIA GPU, you will be using the Pytorch chains. If you are using an AMD or Intel GPU, you will be using the NCNN chains.

• If you are on an AMD or Intel GPU, and the model you downloaded is a .PTH file (Pytorch), you will need to first convert it to NCNN. Here's how to do that.

Upscaling a Single Image

• Based on your GPU, go to the applicable folder (Pytorch or NCNN).
• Download the Simple Image Upscale.chn chain.
• Load the chain into chaiNNer.
• If you've installed chaiNNer correctly, you should see something like the below (the example is from the Pytorch chain).

• (1) load the image you wish to upscale, then (2) load the model you downloaded.
• Click the green button on the top, and begin upscaling!

Upscaling Multiple Images

• Based on your GPU, go to the applicable folder (Pytorch or NCNN).
• Download the Simple Image Iterator.chn chain.
• Load the chain into chaiNNer.
• If you've installed chaiNNer correctly, you should see something like the below (the example is from the Pytorch chain).

• (1) load the image you wish to upscale, then (2) load the model you downloaded.
• Click the green button on the top, and begin upscaling!