Verified: a year ago
SafeTensor
Type | |
Stats | 404 528 34 |
Reviews | (53) |
Published | Feb 10, 2024 |
Base Model | |
Training | Steps: 5,350 Epochs: 1 |
Usage Tips | Clip Skip: 2 |
Trigger Words | mannequin |
Hash | AutoV2 175CC25E7F |
This is somehow not the strangest thing I've made and posted to this site. It's just more on the weird side of things.
What this does is helps to make photos more like a mannequin and less like a doll. Something to note is that a large portion of the training dataset is more realistic-looking mannequins, so they have realistic features as a result. If you don't want that, you could add what you don't like to the negative prompt and try that. Or 30 seconds in GIMP works as well.
For all two of you people who follow me on DeviantArt, I will occasionally use this.