v2.0 Notes:
this is NOT a double trained model so weights of 0.6-0.9 are good
Usage: bajoran, cuff chain earring
Tips: add military uniform, badge to negative prompt to pull away from starfleet uniforms. Also setting things "outdoors" and specifying clothing helps.
Known Issues: The earring chain can get a little weird and get integrated into hair. A few generations or a bit of inpainting may be necessary.
v1.0 Notes:
NOTE: This is a double-trained lora so make sure to use weights < 0.4
If you run this at full strength you are going to have a bad time.
For more information about this training method, refer to this post here: x2 Training Method
This model creates Bajorans from Star Trek
Usage Tips:
Trigger: bajoran (man/woman), cuffed earring, jewelry
Earrings can be a little tricky to get the chains to look right. Some cleanup work might be needed. I'm planning to do another version where I add in specific training images on just the earrings.
Description
Revised training to focus on the earring for more cohesive generation
No longer a double lora so standard weights (0.6-0.9) apply
Focused effort to reduce combadges and starfleet uniforms. Still need to add military uniform, badge to the negative with a high weight. Avoid science-fiction themed prompts helps as well, such as setting it "outdoors".