Okay, so today I messed around with “Mulan Hernandez.” It’s not a person, just FYI, in case you were wondering. I was playing with Stable Diffusion, and it happened to be one of the models/checkpoints that I was using.
First, I opened up Automatic1111, you know, the usual web UI for Stable Diffusion.
Then, I made sure I had the correct Checkpoint/model. The model was already downloaded previously, so I just chose it from the drop-down menu at the top.
Experimenting with Prompts
Started with some basic prompts, just to see what it would spit out. Stuff like:
- “a photo of a woman, Mulan Hernandez style”
- “beautiful woman, portrait, Mulan Hernandez”
I played around with the sampling steps, usually somewhere between 20 and 30. And I kept the CFG Scale pretty standard, around 7.
I generated a bunch of images, just tweaking the prompts slightly each time. I was trying to get a feel for what kind of style this “Mulan Hernandez” produced. Sometimes I’d add things like “highly detailed,” “intricate,” “8k,” you know, the usual buzzwords, other times, I just kept it minimal.
The key is to experiment with various models and prompts until you get a hang of how it works. Once I got the results that I wanted, I stopped there.
That’s prety much it. Remember to save the images you like, other wise you might lose them.
