Generates modern military helmets with mounted night vision goggles, also called NODs (Night Observation Devices).Flux's understanding of a modern military helmet is about 20 years in the past and prompting night vision goggles will result in normal goggles with a green tint.For the uninitiated, amongst some in civilian gun subculture, a helmet equipped with night vision goggles is the ultimate status symbol and a lot of art and lifestyle apparel companies often use this aesthetic.High quality bump (non-ballistic) helmets are anywhere from $300-$400 USD, quality ballistic helmets, like those depicted in this LoRa, are north of $1,000 USD.Binocular NVGs range from $8,000-$12,000 and GPNVGs (QuadNods) are north of $40,000 USD.Quality hearing protection/communication equipment will also cost several hundred USD.Model Tips:All of the images in the showcase are using the LoRa at 1.0 strength, if you are using other LoRas or are using a wordy prompt, I would suggest increasing the strength, I've tried up to 1.25 with good results.Prompting:The keyword, "n0ds" is trained into the LoRa but I haven't had to use ituse the phrase, "wearing a modern military helmet equipped with night vision goggles"it will produce better results if you describe aspects of the helmet and goggles you can reference my showcase images and an example would be "wearing a modern military helmet equipped with night vision goggles The helmet has a camouflage pattern and is equipped with communication equipment, nightvision two tubes with green lenses above her eyeline"For Quadnods, "the night vision goggles are GPNVGs with 4 black tubes with 4 green lenses at the end". You can choose the color of the helmet and lenses. If you are not getting the results you want for these in particular, increase the strength of the LoRa.The LoRa may improve other modern military gear, but it is versatile because I trained it on images where the helmets in non-military contexts, due to the popularity of it being used in the civilian gun culture as a status symbol. I've also used images I've generated on a mostly failed LoRa for the same thing in SDXL. This concept was pretty difficult to train for me.