SimpleTask's Summertime Saga Style [Qwen]

Experimental model trained for roughly 1.5 hours on a 3090. Credit to @luisa_pinguin for sharing their training settings which helped me. This is my first lora so I won't say it's perfect. The dataset could be better but overall I'm okay with this result. I may keep working on it, we'll see.

Some quirks: the eyes can be a bit hard to control due to the fact that most of the content from the game has the characters looking sideways. I tried my best to steer away from it but that's something I could work on later. In addition, you may get weird limbs sometimes (could be due to lack of full body images in the dataset or my constrained training).

Trigger word: clicky cartoon