LoRA from Gundam Hathaway by Gigi Andalucia

Update log

V1.0

First version

V1.0 half

Since there were comments stating that the initial version was overfitted, I am releasing a version trained for half the number of steps—an intermediate checkpoint generated during the original training process. I have no plans to recreate the model unless I personally feel the need for it.

V1.1

Train on a new dataset of 100 images

Please do not make any requests.

Creating a LoRA model is a process that takes at least one day, from building the dataset to adjusting the labels, training the model, and verifying its performance.

I often receive comments like these after releasing a model:

"Please make a character named [Name]."

"Do you plan to make a [Name] version?"

I find receiving comments like this incredibly tiresome.

As stated in my profile, please do not make model requests. Do not ask for new characters or new versions. I am simply sharing the results of a project I pursue for my own enjoyment.

Asking "Do you have plans for a new version?" when I haven't announced anything is essentially demanding a remake. It is extremely presumptuous.

Please don't act as if you are critiquing a meal after someone has kindly shared their packed lunch with you; this isn't a restaurant.

If you aren't satisfied with it, please find something that suits you or create it yourself.