Recreate the HEV power armor from the game "Half-life"
You can see that the auto-detected resource displayed above the image if it was published elsewhere. There won't be this icon if you directly publish it here. Please consider publishing the image directly here （By click“+ Add Post” under this page near Gallery）as it can bring me free buzz,much appreciated！
About this version
this is a clothing lora，good for boy/girl/futanari，recommend use weight of 0.8-1.0
main trigger: HEV_λ,HEV suit from Half-Life,hev suit,hazmat suit, bright orange and dark gray armor, lambda logo on chest, Technological Armor,black gauntlets,orange vambraces,Orange striped paint,gorget,orange codpiece,power armor,boots
special trained for：underground research lab，red Emergency lighting，broken glass volumetric flasks，blood on face，blood on armor
You can refer to the prompts in the example pic.
The training dataset is quite limited, so you can appropriately reduce the weights to incorporate other content you prefer.
Please feel free to experiment and point out my shortcomings. Please share your generations below!
Description
FAQ
Comments (3)
You don't need nearly that many tags for a clothing LoRA, especially when they're not established tags. For a one piece outfit, just "H_E_V" and color would be sufficient. Multiple tags just makes the whole thing weaker and prone to false positives.
Thank you for telling me all this. I really lack experience and skills in clothing lora design. My friend told me that segmented design might help to change the colors of different body parts,My training materials were all low resolution, some were MKIV and some were MKV, and many parts had different colors, so I eventually selected a color for each part.I really don't need that many prompts indeed as you said
@MSTShiBao If there's no fine details, don't worry too much about low res training data as long as you can get a minimum of 256x256. The depthmap training sees an image as isn't all that resolution dependent. I've done a few styles off 480p cartoons and they work fine. I even did a LoRA almost entirely off 108 pixel tall sprites at 4X next nearest just to see what happens and the main issue is it needs a face fix for non-portraits. I previously did a character off even smaller and less detailed sprites and even that's not an an issue due to her bigger and less complex head (though it's more prone to inserting false details on common clothing types without heavy negatives).
Details
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.