This model is trained on 40, 1344x720 images.
Creating a good dataset for this model was quite difficult so don't expect a great result on your first generation. I'm planning to improve this model in future so any recommendations would be great!
I also found Philippe Buchet Style works great with this lora.