Pttrns
Stable Diffusion is known for its ability to generate images through prompts with its txt2img model. However, it extends further by considering an image as a reference and generates an image through its img2img model, making it handy for various purposes.
In this guide, we will present a demonstration featuring a poorly drawn apple and enhance it to make it reliable through the img2img model.
img2img on Stable Diffusion
- Start by opening Stable Diffusion on your device, navigate to the img2img option next to txt2img, and click on it.
- The img2img model comes included with Stable Diffusion WebUI and works with the set checkpoint. After it is loaded, navigate to the Click to Upload text and select the image you want to work with.
- If you do it right, you will find the image uploaded. Once the image is uploaded, type in the positive and negative prompts.
- Before you proceed further, scroll down to the bottom and adjust the settings as set in the image below. If you have specific requirements, feel free to set them accordingly.
- Once done, click on the Generate button and wait for the image to generate. Now, you will find an enhanced image generated to the right according to the prompt provided.
Here is the comparison of the image before and after enhancing it on Stable Diffusion:
Wrapping Up
Congratulations on learning how to use the img2img model on Stable Diffusion!
Stable Diffusion generates reliable results if you get the prompt and other factors right, but inconsistencies are commonly found. For instance, Stable Diffusion generated an image with an apple that had two stems and required editing.
If you’re not happy with the results, adjust the CFG scale and Denoising strength sliders until you get the desired result. However, if nothing works, try changing the checkpoints and experiment with different models.
Related Reading
What Is a Negative Prompt in Stable Diffusion
