I want to make a rainbow lion. This is the lion as made by the prompt alone, with no initial image. It has rainbow fur, but I want a different pose.
Unsplash has many beautiful public domain photos which are perfect for initial images. I choose one with good colors, nice composition and no distracting elements:
Adding that initial image gave me this image:
With the prompt strength at 0.80, the AI is using the colors from the initial image, but not the pose. So I reduce the Prompt Strength.
When I dial it down to 0.70, the pose from the initial image is incorporated into the final artwork.