This version was re-trained using HD materials. It fixes the issue from the previous version where it only generated children. This version is well-balanced, generating both adults and children properly. Although this version also does not require a trigger word, since it focuses on smile optimization, it is recommended to describe facial expressions for better results.

Prompt Reference:

There is a childlike smile on their face. (Childlike smile on the face)

There is a shy smile on his/her face. (Shy smile on the face)

Laugh out loud with an open mouth. (Laughing out loud with an open mouth)

In short, make sure to at least describe the expression to get the desired effect.

Additionally, I'd like to share my method for writing prompts.

For example, if I don't know what outfit the character should wear, I search online for some outfit combinations that look good to me, send the image to an AI, and ask it to write the description for me. Here, I'm using DS as an example.

The same logic applies to other situations. For instance, if you want to replicate an image without using image-to-image, you can simply send it to an AI, ask it to write a prompt with a similar vibe, make a few tweaks, and translate it. FLUX's advantage lies in natural language.

Finally, I hope everyone creates awesome images that truly catch the eye!

To be straightforward, I'd love to see your generated image posts and get your likes ❤

Reposted from: Say No to Stiff AI Expressions, More Vivid Expressions