To vary a bit from cinema i've trained this lora on site, the dataset is made of 900 or so midjourney images, carefully curated, unfortunately the caption on site is just poor and i think this might be droping the concept understanding of this lora.

However this is still an enjoyable lora to use, it was mainly made for photography but has been trained on illustration and comics a bit,
excited to see what you guys will produce with this

Due to long render time i haven't been able to properly test it, but i'd say the best weigh is around 0.7 , you have some wiggle room between 0.5 to 1