Type | |
Stats | 592 1,117 |
Reviews | (89) |
Published | Jan 14, 2024 |
Base Model | |
Usage Tips | Clip Skip: 2 |
Trigger Words | yomama_beam |
Hash | AutoV2 FFFFA50698 |
Beam pose
First off, belated happy new year to everyone! Thanks to a financial support, I am able to commit in making this update, and this update was allowed to be released publicly generously, so thank you.
This lora works better horizontally, also work well alone at weight 0.8; 0.9 if you are pairing it with another character's lora, drop the weight of the character's lora to about 0.5-0.7 for better results.
V1.0
updated with generated images, results can still be varied between checkpoints, but success rate increased in general, less artstyle bleeding observed. Adjust the weight and sampling method to test what work best for you. This lora currently still work best for human only at this moment, if you wish to change to a non-human character, inpaint may be required.
If you want to mix with a character Lora, I suggested you lower the weight of the character to 0.3-0.5 in txt2img first, increased it higher later in img2img.
Added tags for actions
Charging: posture of holding energy ball next to hip
Shoot: posture of shooting out, both hands in front of chest (rather unstable)
Beam: posture of shooting out with beam blasting out, from side or from behind (very unstable)
Suggested tags for negative prompts
(dougi), spike hair, ((wristband))
V0.5
This Lora can teach your character how to perform kamehameha, this is still a bit unstable, sucess rate is about 40~70% depending on the checkpoints. Weight working rage from 0.8-1.2; this lora work alone best at 0.9.
The poses can be broken sometimes, i suggest that you find a pose generated you like and use that in controlNet (or use my displayed image for the pose).
If you don't see any energy ball when you pair with another character's lora, adjust this kamehaemeha lora to higher weight (0.9-1.2). img2img and inpaint may be needed depending on the other character's lora. The prompt in the display gallery will only get you general composition, further editing is needed in most cases.
Feel free to adjust the prompt to get the results; such as ((energy ball)) or (electricity) to enhance the effect. I suggest including night scene in prompt so the visual effect can stands out.