Beverly Crusher Badass (Dreambooth Model) Star Trek Fan Art

Name: Beverly Crusher Badass (Dreambooth Model) Star Trek Fan Art
Rating: 5 (15 reviews)
Author: user9

312

Updated: Nov 10, 2023

character

star trek female fanart dreambooth beverly crusher

Download (5.57 GB)

Verified: a year ago

SafeTensor

This checkpoint includes a config file, download and place it along side the checkpoint.

Details

Type	Checkpoint Trained
Stats	312
Reviews	Positive (30)
Published	Nov 10, 2023
Base Model	SD 1.5
Training	Steps: 1,000
Trigger Words	40 year old woman closeup rear view side view full body shot
Hash	AutoV2 91A6D965B2

2 Files

default creator card background decoration

user9

License:

CreativeML Open RAIL-M Addendum

This resource is intended to reproduce the likeness of a real person. Out of respect for this individual and in accordance with our Content Rules, only work-safe images and non-commercial use is permitted.

If you are this person or a legal representative of this person, and would like to request the removal of this resource, you can do so here.

EDIT: THIS IS AN ENTIRE CHECKPOINT MODEL DEDICATED TO THIS CHARACTER NOT A LORA. ADD YOUR STYLIZED LORAS AFTER LOADING THE MODEL, IT WORKS BETTER THAT WAY. THIS WAS TRAINED WITH DREAMBOOTH.

This is Dr. Beverly Crusher from Star Trek the next generation. You can run this model at 512x512 (recommended) or 768x768. I would recommend using Adetailer extension for cleaning up more distant face and hand shots. This model was Dreambooth trained with Koyha-Gui on 128 images, closeups, distant shots, merged with other models, then re-merged again at very low strength 1.0 with one of the "overtrained" checkpoints Kohya spit out at the end. If you have questions on the worldflow please feel free to ask. Dreambooth definitely gives better results than SD 1.5 Loras and SDXL Loras for subjects.

I recommend using --no-half --no-half-vae for startup argument in auto1111 for higher quality, and the VAE vae-ft-mse-840000-ema-pruned.safetensors for this, adding "weird eyes" to the negative prompt. You can also use terms such as "full body shot" for a more distant view, "rear view" "side view" "closeup"