Trained on 25 images of Jenna Ortega, 1600 steps :D
v1.1 improved quality, no trigger required. This is the same model I posted on huggingface a few days ago.
v1.0 does not require a trigger word
Description
FAQ
Comments (10)
Hey, Cool lora, but for some reason it does not seem to work in my setup, do you know if its because im using the fp8 version's of Flux and T5? thanks
i even downloaded your images and tired the exact same workflow.
i have not tested them with the fp8 versions, D:
update your comfyUI, they fixed the flux Lora issue
@razzz i'm using forge, i'm assuming it's not fixed there yet?
@Stagnation as far as i know, flux loras do not work with forge just yet
any chance to train it without having to use a trigger or at least the name spelled normally.
After further testing, this model does not seem to require a trigger word. The trigger "j3nnaort3ga" may help it but it generates an image of her regardless. I’m still figuring out how SimpleTuner handles trigger words.
Thanks for your released lora's so far. I'm trying a run as well soon, what learning rate did you use for training a person or do you feel gives the best results so far?
hii, the learning rate i use is 5e-4, which works really well. if you would like to see all my settings, check the guide below ^^
https://civitai.com/articles/6678/
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.