This LoRA adds a scifi-ish helmet with head-mounted display covering eyes - one of concepts many PDXL models were always struggling with, especially if paired with character. Works pretty okay with character LoRA, not much with innate PDXL characters.<h3 id="v.0.92-info:-5kuc3qaku">v.0.92 info:</h3>Trigger word changed to stdt001, combined with other tokens:helmet gives a distinctive helmet look;covered eyes extends it down and covers eyesvr headset adds an "embedded" virtual reality visor to the helmetRecommended strength varies between 0.8 and 1.2, depending on chekpoint and additional LoRAs.The helmet structure seems to be more consistent now, albeit it still works better with semi-realistic checkpoints than anime or photorealistic ones.<h3 id="v.0.9-info-below:-63hx2jhmw">v.0.9 info below:</h3>(Temporary) trigger word is "v1sorm4sk" but I'll probably change it in v1.0 because it clashes with v_sign and results in victory gesture appearing.Use helmet and/or head-mounted display to control features, adding science fiction, covered eyes and/or cyberpunk might also help. Recommended weight is ~1.0.Add realistic in negative prompt if you're using it on anime checkpoints and realistic styling is leaking too much. also this is my first time cooking a LoRA for PonyXL! Still many things to learn and obviously it struggles with all but portrait poses but I actually like the outcome, so v1.0 and/or IllustrousXL version might come later =)

VisorMask [PDXL]

same dataset but better retrained =)Trigger word changed to stdt001, combined with other tokens:helmet gives a distinctive helmet look;covered eyes extends it down and covers eyesvr headset adds an "embedded" virtual reality visor to the helmetRecommended strength varies between 0.8 and 1.2, depending on chekpoint and additional LoRAs.The helmet structure seems to be more consistent now, albeit it still works better with semi-realistic checkpoints than anime or photorealistic ones.

Files with SHA256: 7401a4533465e927cdea1ce8b80b89926d13a2ff0943af501268d5786d92307a