Type | |
Stats | 6,017 |
Reviews | |
Published | Apr 28, 2023 |
Base Model | |
Trigger Words | oversized_shirt |
Hash | AutoV2 8A5C8A6643 |
Get early access to build and test build, be able to try all epochs and test them by yourself on Patreon or contact me for support on Discord
To support me you can buy me a coffe at https://ko-fi.com/antonioriolo
Training settings:
36 images, 20 steps, 13 epochs, you find many epochs (with test images using plot xy script) on my Patreon while supporting me.
Recommended settings:
Make sure you are aware on the usage instructions of LORA
Sampler: DPM++ SDE Karras (Recommended for best quality, you may try other samplers)
CFG Scale : 5 to 9
LORA weight for txt2img: reccomended 0.2-0.7 (i used 0.5 in my images).
Keys: oversized_shirt
Somehow i couldn't manage to select the color of the shirt or to have sports or colorful shirts, most of the times it just generate a white empty shirt but there is not a single one empty white shirt in the training model so i really don't know. maybe it just need an higher weight. If you find a good prompt post it for the others please, thanks in advance.
In my images i gave a weight of 1.3 to the keyword oversized_shirt, and i placed the keyword at the beginning of the prompt.
There are a few images in the training with "back_view" keyword but it doesn't always work, you might try it and give an higher weight if necessary.
There are also many images with "colorful" keyword, you might try it with higher possible weight and hope for a lucky seed.
I posted epoch 10, maybe with other epochs all these problems disappear, if you want to try all epochs come in my patreon and get Edge Access.
Checkpoint used to generate the images: Chillout for realistic ones/Yesmix for anime ones
Enjoy!