Update hypernet for the 1.5 model that greatly improves overall portrait quality

Adjusted folder structure for better clarity with the release of SD1.5 and added first version of starsectorportrait model trained on SD1.5

Added a v2 of the finetuning hypernetwork which was more aggresively trained and seems to yield higher quality images

added finetuning hypernetwork trained on main dataset without face covering helmets