This LoRa was trained on artwork from Nobunaga's Ambition: Sphere of Influence. Unfortunately, most of the images used for training possessed a lot of empty black space that I couldn't be bothered to remove, and this had passed onto images generated by this Lora when used alone. If you want nice, clean images like the examples, use an additional Lora together with this one. ^Thus^ I recommend using this Lora at 0.5 strength in tandem with something with a similar styling. As you can reasonably expect, this LoRa is best at generating imagery from Feudal Japan. Unfortunately, much of the training images had a lot of blank black space, and that will come about in image generations sometimes. It's important to include 'empty space' and 'black bars' in the negative prompt to ensure a full image is generated, though that won't always work. As such, I honestly recommend this LoRa be used in conjunction with another, it's excellent when acting in compliment to other styles.