FLUX retrain: <a target="_blank" rel="ugc" href="https://civarchive.com/models/930430">https://civarchive.com/models/930430</a> V2 UPDATE: Expanded the Dataset and used higher resolution images<a target="_blank" rel="ugc" href="https://pastebin.com/E8pJTFw1">old captions work, V2 captions here or use Training Data link</a>(few terminids and zero automatons in training set)Style model, Trained with 11 images using CivitAI on site Lora Trainer, using default settings. Caption Root:<ul><li>hll style, close up helldiver helmet</li><li>hll style, helldivers with long capes and helmets team posing, holding big gun</li><li>hll style, helldiver standing with big gun, in front of helldivers raising flag</li><li>hll style, low angle, helldiver standing, holding big gun</li><li>hll style, helldiver outfit with helmet, fist salute</li><li>hll style, helldivers posing with big guns</li><li>hll style, helldivers posing with big guns</li><li>hll style, helldiver outfit with long cape and helmet, holding big gun</li><li>hll style, helldivers are raising a flag</li><li>hll style, helldivers are standing together, holding big guns</li></ul> The Caption Root was supplied with the images to the CivitAI trainer, then using Append we allowed merging with autotagging, to expand the token range. Caption Merger:<ul><li>hll style, helldiver outfit with helmet, fist salute, solo, man, armor, male focus, helmet, desert, mountain, belt</li><li>hll style, helldivers with long capes and helmets team posing, holding big gun, weapon, multiple men, gun, science fiction, cloud, armor, rifle, cape, sky, holding gun</li><li>hll style, helldiver standing with big gun, in front of helldivers raising flag, weapon, multiple men, armor, gun, cape, rifle, helmet, holding weapon, holding, sky</li><li>hll style, low angle, helldiver standing, holding big gun, weapon, gun, solo, helmet, man, rifle, assault rifle, armor, male focus, holding</li><li>hll style, helldiver outfit with long cape and helmet, holding big gun, weapon, armor, solo, gun, man, helmet, cape, male focus, rifle, holding weapon</li><li>hll style, helldivers are raising a flag, multiple men, armor, cape, flag, weapon, helmet, sky, outdoors, desert, gauntlets</li><li>hll style, helldivers posing with big gun, weapon, multiple men, gun, rifle, cape, armor, holding weapon, science fiction, holding, 2men</li><li>hll style, helldivers posing with big gun, weapon, gun, multiple men, armor, cloud, helmet, science fiction, 2men, cape, rifle</li><li>hll style, helldivers are standing together, holding big guns, weapon, armor, gun, multiple men, cape, rifle, science fiction, helmet, spacecraft, energy gun</li></ul>designed for use with reference images and IPAdaptorV2, although you can use whatever system you prefer :) ComfyUI users must ensure to use the SDXL Clip encoder, if you use the normal one, Clip L will not be loaded, resulting in a partially formed style. (This is normal, because it allows a workflow to be used for 1.5 and 2.1 without changes, where SDXL used two clip models)

Files with SHA256: 943b9f43a81c7024140f33fba7f139fcf30ece8a4512e7a9680fdbd412500ced