⚡️ Buzz Girl ⚡️

⚡️ Buzz Girl ⚡️ FLUX v2

Finally, the complete set. Whereas 1.2 was just her body and face, Version 2 was trained on 99 images with different clothing and various expressions and poses. Both of her suits (main and black stealthy suit), her Courtney Kwent look, and various casual and sporting attire.

• You will no longer need BuzzSuit lora to call on her costumes. The trigger phrases are in the trigger section for ease of use. If you put just "BGsuit" or "BBGsuit" without the rest, it does some crazy variations. Or try "BGsuit: a yellow and black superhero suit with blue trim" or other variations... and it does some neat stuff.

• Every caption was done manually and then tested, adjusted, and tested again. I am confident this is by far the best version of Buzz Girl thus far.

• Dataset images were made from using v1.2 and the BuzzSuit lora as well as a few good images from earlier versions.

• Many thanks to @Heim777 who made a few of the images used and also played a big role in thinking up the character, her look, and lore.

• NSFW base model is still in there. So give her something to wear!

• There are some animated style images in there, so I am curious how those kind of prompts will work.

• Now with more rescued puppies, kittens saved from trees, and possibly raccoons freed from heavy dumpster lids!

• The most time consuming part was generating hundreds of images and selecting only the best ones.

• SDXL version coming sooner or later (depending on how well the tagging takes)

ENJOY! And thank you to everyone who has supported the Buzz Girl saga.

v1.2 ⚡ Buzz Girl — Character Foundation & Shape

Buzz Girl is a bright, energetic superhero who lives in the high-tech media city of Buzz City, working undercover as journalist Courtney Kwent for the Daily Challenge News (DCN).
She’s kind, clever, and always on the move — rollerblading through the city, rescuing cats, and exposing corruption with a smile and a spark. Buzz Girl is in her early 20's.

This is version 1.2 of the official Buzz Girl LoRA, which builds on multiple beta models to provide a clean, accurate, and anatomically stable base for Buzz Girl’s full character range — especially her BuzzSuit superhero overlay.

📦 Description:

This FLUX LoRA provides a clean, unclothed anatomical base for Buzz Girl to support accurate and distortion-free clothing overlays (such as her iconic BuzzSuit).
It is not intended for NSFW use. The unclothed state exists only to help clothing layers match her form correctly during generation. (Creator's comment: That being said, she is a fictional character made entirely from AI and words. I'm not here to judge.)

The model was carefully trained to retain:

• Buzz Girl’s core identity and body shape

• Realistic anatomy and limb structure

• Compatibility with superhero costume overlays (and other clothing loras).

⚙️ Training Details:

• Images: 48 instance images (×4 repeats), 27 regularization images

• Resolution: 1024×1024

• LoRA type: FLUX, 32 dim / 24 alpha

• Steps: 1,536

• Epochs: 4

• Noise offset: 0.1 (to smooth rib and skin detail)

• Regularization: Yes — uses general “young woman” examples for realism balance

🎯 Usage:

This model is intended as a foundation LoRA — best used in combination with:

• ✅ The BuzzSuit LoRA (superhero costume overlay)

• ✅ Hair, eye, or expression LoRAs

• ✅ Style, pose, or environment enhancements

To dress her up, simply apply clothing LoRAs like BuzzSuit over this base with the appropriate LoRA weight. Or specify the clothing wanted in the prompt.

Special:

This Lora should also work as a NSFW Unlock of FLUX, but you will only get Buzz Girl's look out of it. At least she isn't too hard on the eyes.

🛡 Disclaimer:

This model was built strictly for character fidelity and artistic use.
All images were AI-generated by the creator and captioned manually to support clean, consistent anatomy for character-building.

This LoRA is not intended to resemble any real person, living or deceased.
No real photos, celebrity likenesses, or copyrighted content were used.
The character was created solely through prompted image generation, using a combination of:
– Custom character prompts
– AI-generated concept art
– Manually captioned dataset examples