This is a rank 64 Chroma1-HD lora (should work on Chroma1-HD-Flash too), trained in over 500 images for 68 epochs.
This was trained with the goal of fixing some of Chroma's anatomy issues (eg characters holding objects such as swords, cigarettes, lollypops, guns, bows, and other things), and improve aesthetics for fancy styles such as retro photography, CGI art and others.
Unfortunately the goal was not completely achieved at this time, but consider this model a work in progress, and newer versions of it could be made.
Use this as negative prompt, otherwise you may get bad outputs:
