This is a Lora trained on 65 images, with a LoRA ranking of 16 on a RTX 4090, and 128gb of system ram locally. This LoRA is very flexible with realistic images, or the graphics style of the Marvel Midnight Suns video game.
The data set resolution of all images were 3440x1440 ultrawide and given its own bucket then trained at 1024x1024 for best overall detail retention. This LoRA may struggle with different clothes because I specifically tried to keep her super suit consistent across generations.

If you enjoy my work show some love, and share your generations here for me to see!

Please consider giving me a follow, like, and share your creations here! This would greatly help me grow a audience for my work. Happy to be apart of the community here!

This is my first post on here, been training my own LoRA's for about a year now. If you'd like a custom Flux LoRA trained, or any other requests please feel free to message me. I offer making exclusive LoRA's just for you as a service, if you don't mention that you'd like to commission me then I'll write it down as a request.

There is no trigger word just say "captain marvel" instead of "woman" in your prompts.

Recommended weight is 0.7-1.20 any further passed 1.20 will be deformed.

Recommended Guidance: 3.5-4.0 for more constancy, feel free to experiment though.

Recommended Sampler/Scheduler combinations.
Euler - Simple

Euler-Beta

Euler Ancestral- Simple
Deis-Simple

Deis-Beta

Deis-Karras (Will be either very good, or very mid.)
All the dpm options adds different styles/details so start with dpm++2m-Simple and experiment from there.