anima
The dataset includes some illustrations from the game
Art style features: Tendency towards pseudo-thick painting
lora Strength: Default is 1, can appropriately increase the strength，might be a surprise
il
This is my first time training the Lora style. Unlike character Lora, I'm still learning about style training.
This Lora dataset only contains about 100 images. Since it's my first attempt, I didn't set a very high step value.
Some parameters may not have been filled in reasonably during the training, and there is still much room for improvement in the training results.
Here are the suggestions
Trigger words: hbrstyle
Although I added Trigger words during my training, the difference in the results when using or not using Trigger words wasn't very significant after the test
Weights: 0.5-1
For the specific differences between Step, Lora Weights, and Checkpoint, please refer to the comparison chart of the X/Y/Z axes in the example