Trained a bit further for a few epoch after some feedback, it has better style fidelity at the cost of less versatility with other lora, or style prompting. It's not fundamentally better.