This is my first shared LORA and I am still very much in the testing phase and in the process of finetuning my dataset. It does need some work still and I plan on doing it. My aim for this LORA is for it to be an all-in-one Fallout Attire LORA complete with attire from all factions. Starting with Vault Dwellers.

PLEASE SHARE YOUR CREATIONS!

Trained on 50 images of cosplayers and in-game screens with hand written captions.

Ive mostly tested it with Colossus Project XL and Better Than Words but it should work fine with any XL checkpoint.

For the most part it plays well with other LORAS

Recommended Strength: 0.5-0.76

It sort of understands what a Pipboy is, though in hindsight I should have provided some better images of a Pipboy in my dataset. Ive had okay results removing the Pipboy altogether by putting Pipboy in my negative prompt. Alternatively, you can add "wearing a Pipboy on his/her wrist" to try and generate a Pipboy.

Example prompts:

"A woman/man wearing a Vault Suit"

"A woman/man wearing a Vault Suit with Leather Armor"

"A woman/man wearing a Vault Suit with Leather Armor and a Pipboy on his/her wrist"

Adding "holding a gun" may also produce a gun but only a few images in the dataset featured a gun.

Adding "wearing a Vault Suit with the number 76 on the back" works with very mixed results. You can input whatever number but 76 and 111 work the best as that is what is present in the dataset.

I highly recommend prompting your background such as "in a desert wasteland" or "in an abandoned power plant" etc.

Adding specifics like hair color, eye color, etc. can help with Face LORAs.