Download
1 variant available
5510 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
18.5K0 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
2.4K0 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
(80)
Sep 4, 2024
Trained on 96 images.
I haven't had much time to test or expand on this one, but I've enjoyed the results so far.
Sharing as an alpha version for now, I hope to clean up the concept tagging for better control and expand the dataset a bit more for further training.
Show more

6.5K0 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9K
11.2M0 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9.0 1 2 3 4 5 6 7 8 9M

License:
CreativeML Open RAIL++-MEye Focus Close-up
This is a LoRA trained on a small collection of eye-focus close-up's.
It's intended for use with inpainting using the adetailer extension - any YOLO model segmenting faces/eyes/heads should work nicely with it.
It can also be used as-is for upper body or face close-ups.
Recommended prompt:
Images were manually rated with quality/aesthetic tags:
masterpiece, best quality, low quality, very aesthetic, aesthetic.
Using something like this in your adetailer prompt, or the main prompt for a close-up is recommended:
Positive prompt:
eye focus, close-up, {color} eyes,
{tags},
score_9,score_8_up, score_7_up, score_6_up, masterpiece, best quality, very aestheticNegative prompt:
(worst quality, low quality, sketch:1.1), score_4, score_3, score_2, score_1, error, bad anatomy, bad hands, watermark, ugly, distorted, signature

